Essays where futures research gets less vague.

Futures studies has many methods: scenarios, Delphi studies, trend analyses, simulations, backcasting, roadmaps, weak signals, Causal Layered Analysis and many more. Usually, we call this a toolbox. A researcher chooses the correct tool, uses it for the problem and produces some knowledge or insight about futures.

But this description is a little too clean.

A method is never only a tool. Every method comes into the room carrying assumptions about causation, evidence, authority and purpose. It assumes why the present can tell something about the future, what counts as evidence, whose judgment should be trusted and what the operation is trying to achieve.

These assumptions are rarely written down. They operate under a kind of methodological omertà: the family conducts its business, but the rules behind the business are not spoken aloud. To make them visible, let us arrange the methods into five families and call a sit-down.

The Exploratory Family

At the first table sits Don System Dynamics. Trend Analysis serves as the consigliere, while simulations and extrapolations form the crews. This family controls the territory nearest to the present. It asks: If the present system keeps doing its business, where will it take us?

Its evidence is historical, quantitative and aggregated. Its strength is showing feedback loops, delays and cumulative effects that intuition misses. Its weakness is continuity. The future does not have to copy the past, but the mechanisms producing the past must remain stable enough. A new technology, political transformation or cultural break may enter the neighbourhood without asking permission.

The clock keeps ticking until the spring snaps.

The Expert-Judgment Family

At the second table sits Don Delphi. The Expert Panel is the underboss, structured interviews work as capos, and specialists are soldiers and associates. This family believes that important knowledge about futures sits tacitly inside expert judgment. Its job is to make this knowledge talk.

Delphi uses anonymity, rounds and controlled feedback so one powerful person does not take over the room. It is useful when formal data is weak or knowledge is spread across specialties. Still, agreement is not truth. Similar experts may share the same education, interests and blind spots, while a numerical average can make uncertain judgment look more exact than it is.

Sometimes confidence puts on an expensive suit and introduces itself as probability.

The Plural-Futures Family

At the third table, Don Scenario Planning refuses one official future. Morphological Analysis serves as consigliere, while Cross-Impact Analysis checks whether the stories contradict each other too badly. This family believes that the future comes in plural. One scenario receives meaning from its place among alternatives.

Good scenario work disciplines imagination. It makes people consider developments they dislike, do not expect or find uncomfortable. Evidence supports the assumptions and relations inside a scenario; it does not prove the scenario as a prediction.

But the family has its rackets. Somebody may become attached to one dramatic or politically useful pet future. The scenario is selected first, and the evidence is brought in later as an associate expected to support the boss. Sometimes four futures are put on the wall, but everybody knows which one is treated as the real future.

The Normative Family

At the fourth table sits Don Backcasting. Roadmapping is the consigliere, visioning exercises are the capos, and workshops provide the crews. This family begins from a desirable future and works backward. It asks: Where do we want to arrive, and what actions could take us there?

The family honestly recognizes that futures work involves values. Data cannot decide what counts as justice, sustainability or improvement. Evidence can show barriers, costs and necessary steps, but it cannot choose the destination.

The danger is that the destination is fixed before a proper sit-down. A small group can decide what future everybody should desire and then present the roadmap as technical work. The route may be excellent, but perhaps nobody outside the family agreed to make the trip.

The Critical-Interpretive Family

At the fifth table sits Don Causal Layered Analysis. Narrative, discourse and metaphor approaches move through the deeper levels of the organisation. This family looks below visible events toward social structures, worldviews, myths and metaphors.

Some futures appear realistic because they fit the dominant worldview. Others stay outside imagination because we lack concepts and stories for them. The family’s evidence is language, images, narratives and shared assumptions. Its strength is showing that an apparently neutral future may already be doing business for a particular worldview.

Its weakness is that interpretation may never end. Dig deep enough and every plan looks like ideology, every model like propaganda and every scenario like a cover operation. But eventually somebody has to leave the back room and act.

The Commission

None of the five families is simply wrong. Each controls a different territory in the possibility space. The Exploratory Family sees continuity. The Expert-Judgment Family sees tacit knowledge. The Plural-Futures Family sees alternatives. The Normative Family sees values and direction. The Critical-Interpretive Family sees the frames deciding what can be imagined.

Each family gains its power by accepting assumptions. These assumptions are the price of doing business. Problems begin when methods cooperate without discussing the terms. A model places authority in equations and model-builders. Delphi places it in expert judgment. A workshop may place it in participation. A critical method may question the categories used by all the others.

Such cooperation can be productive, but it needs something like the Commission: not one supreme method ruling the city, but a place where territories, assumptions and conflicts are made explicit. Without this, methodological diversity may be only confusion with a nice deliverable.

This is where philosophy can help futures studies. It should not arrive as an outsider demanding the whole city. Its role is closer to a careful consigliere: What does the method assume? What evidence does it recognize? Who has authority? Which possibilities does it see, and which disappear? What is the operation trying to achieve?

The goal is not to eliminate the five families. Futures studies would become poorer if one family defeated all others. The goal is to understand who is sitting at the table, what business each family conducts and what each quietly expects from the rest.

Because every method has its loyalties.

And every method has a price.

Introduction

In the early 1970s, Pierre Wack, Ted Newland, and their colleagues in the Group Planning department at Royal Dutch Shell sat down and imagined something that most of the actors in the oil industry considered far-fetched: a future in which nations producing oil restrict supply. The exercise was not guessing, prophecy, or what have you. It was what could be called disciplined imagination, foresight even – it was a structured attempt to take a remote possibility seriously and to think through what this possibility of restrictions, in this case, would mean and how one could respond if the possibility was to become reality. Within months, the 1973 oil crisis made that possibility real – it made the possible actual. Shell was better prepared than its rivals. But not because it had guessed right or even predicted the future, but because it had paid attention to, and rehearsed, an alternative one (Wack, 1985a, 1985b; Andersson, 2020).

Wack’s exercise rested on an assumption so basic that it is usually taken for granted: that there genuinely are alternative futures that are worth exploring and that we can make decisions about. This premise is a foundation of the entire field of futures studies as we know it, at least when we turn to wider issues than prediction. Early on, Roy Amara built the assumption within the three commitments that have become central in futures studies: the future is not predetermined, the future is not predictable, and future outcomes can be influenced by present choices (Amara, 1981). Among others, Wendell Bell built an entire epistemological framework on the idea that useful knowledge about the future can be gained (Bell, 1997a, 1997b). Joseph Voros’s widely used version of the futures cone (see Voros, 2003) with its ever-expanding sections of possible, plausible, probable, and preferable futures (building on Hancock and Bezold’s 1994 earlier version) has become perhaps the most widely taught visualisation in the field. It also carries the assumption that there is a set of futures that one needs to pay attention to – an ever-expanding set. Earlier, Fred Polak argued that a civilisation’s capacity to generate positive images (in plural, again) of the future is the primary driver of the vitality of the civilisation (Polak, 1973).

While we take such ideas, about possible and alternative futures for granted, these are not merely working assumptions. Rather, they are, in fact, philosophical commitments, and they carry considerably far more weight than the people in the field of futures studies recognise. To see this, we need to consider only the idea that, perhaps, the future is already fixed. Now, if only one course of events is genuinely possible, then the attempt to study alternative futures may be a practically working illusion that is, deep down, incoherent – a smokescreen to avoid us confessing we are prisoners in time. However, if the number of possibilities is infinite and unconstrained, the field’s talk about alternative futures is, again, a smokescreen to ignore most futures and focus on a limited set of them – while claiming we expand our understanding of agency and so on. Between these extremes lies a rich and contested set of philosophical ideas, a set that futures studies has largely left underdiscussed even as the practicing of futures studies depends upon the resources of these philosophical ideas at every turn.

This text maps those ideas. It proceeds by discussing five questions, each of which turns out to be surprisingly consequential for everyday foresight practice despite their philosophical nature.

Does the future exist at all?

Can the future be genuinely open, or does logic force us toward determinism?

What gives formal substance to the intuitive idea of branching futures?

How wide is the space of possibility?

How do we know what is possible?

Our aim is not, and it cannot be, to settle philosophical debates. Rather, we attempt to show that the debates matter in futures studies; that every time a practitioner draws a scenario tree, judges a future “plausible”, or maybe dismisses one as “impossible”, philosophical commitments about time, truth, and possibility are being made, whether explicitly or not – most of the time implicitly. Making those commitments visible is the first step toward making futures studies able to discuss its ontological foundations, its deepest questions.

Does the Future Exist? Three Ontologies of Time

The most fundamental question for any discipline concerned with the future is deceptively simple: Is the future real? Not real in the sense of “likely to affect us, once, or if, we get there” – everyone agrees on that – but real in the deepest ontological sense: does it exist? Or what does it mean to say that the future exists? The question can be framed in many ways, for example, are future events already out there (already if not now?), waiting to become the present – in the same way a city exists before a traveler to that city arrives? Or is the future genuinely absent, nothing – something that does not yet exist in any sense at all? We are not taking the easy route, but as the philosophy of time offers three basic answers to such questions, we need to turn to that field now. Each of the three answers defines different conceptual foundations for futures studies – even with different practical consequences, odd as that may sound.

Presentism: only the present is real

The first answer is the most radical. Presentism holds that only the present moment exists. The past is gone; the future is not yet here. Neither future nor past has any place in the inventory of what is real (Markosian, 2004). But we say things (true or not) like “Oil price will double” and “Titanic was a mighty ship” – what can presentism say these claims really mean?

On a formal level, A. N. Prior built a logic (1953; also 1967) that, at least in principle, makes this position workable. The standard move in modal-temporal logic would be to treat “was” and “will” as quantifiers ranging over past and future things, the way “some” ranges over present ones – “some things are red”. Prior looked at the issue differently. He treated tenses (“will” and “was”) as primitive logical operators – they are like stamps was and will that you place on present-tense sentences, without needing past or future objects to exist for the stamped sentence to make sense. Take the sentence “a Mars colony houses a million people” – false right now. Stamp will on it, and you get will(a Mars colony houses a million people), which can be true or false today, in the present, while we still do not have to commit ourselves to the existence of any world or state of affairs in which the inhabiting of Mars has actually occurred (or failed, or not attempted, or what have you). The technical name for the stamps is tense operators, and they give the presentist exactly what futures studies needs from a logic of time (“presentist committed futures studies”): a way to say things about the future without quietly populating the universe with futures that are not yet there. On Prior’s picture, this is what futurists are doing every time they speak of the future – stamping the present.

For foresight practitioners, prima facie, presentism provides the strongest philosophical backing for treating the future as open, non-actual, and something we can create. The future, looked at in this way, is not a distant country waiting to be mapped with some cross-temporal telescope. It is a genuinely absent domain that must be imagined, constructed, and chosen into being. This resonates deeply with the activist orientation that people like Gaston Berger brought to futures thinking when they suggested that the future is not something to be passively awaited or mapped but something to be actively shaped through present decisions (Cournand & Lévy, 1973). Presentism can also be seen as a natural philosophical partner for Polak’s insight that images of the future actively constitute the future they describe (Polak, 1973). Creating a statue from marble is one thing; it is another thing to wait for one to be brought to you while waiting to see how it will look, once it comes. Being an artist and reading catalogues are two different activities.

But presentism comes with costs that matter for foresight. If no future this or that exists, what exactly are we talking about when we talk about “the future”? What makes a future-directed statement true or false – and, at the same time, what makes one more justified than another? A claim about the DNA of Donald Duck cannot be more or less justified, or true, because he does not exist. Is this true about the claims about the future too? For example, when Amara says “the future is not predetermined”, what is the “future” he is referring to, literally speaking? If future-directed statements lack determinate truth-values and truth-makers (i.e., are not true or false and there is nothing that makes one true or false), how should we understand the knowledge that futures studies claims to produce? These are not idle, meaningless puzzles. They connect directly to the challenge of evaluating scenario plausibility – of saying that some futures are more credible than others – and they lead us straight to the classical, heavily consequential, sea-battle problem we discuss in the next section.

Eternalism: the future is already there

The second answer is the opposite position. Eternalism maintains that past, present, and future objects all exist, differing only in their temporal location. Consider London and Tokyo. Both exist; the difference in spatial location does not affect what is real. In eternalism, the universe is a four-dimensional manifold – a “block” – in which events occupy different positions; some in different places, some in different times. A shirt changing color simply consists in the shirt being white at one temporal location and pink at a later one. There is no objective “now” that “moves” through the manifold; the distinction between past, present, and future is perspectival rather than metaphysical, like the difference between “here” and “there” (Sider, 2001). Sometimes when I have tried to explain how it can be real that I am 15, I say “there I am”. Then my mother asks “where?” and I answer “There, in the past, in year Y”. Once we do not get stuck in the present issues, we can understand that the present just happens to be now – but that does not make it more real, ontologically, than other moments.

Eternalism draws considerable support from modern physics. Physics, specifically special relativity, eliminates a frame-independent present. This elimination of a frame-independent present – the fact that observers in different states of motion disagree about which distant events are simultaneous – seems to damage any theory that requires an objective “now”, including the presentism we previously discussed and the growing block (see below) (Putnam, 1967).

For futures studies, eternalism presents a paradox. On one hand, it implies that the future has full ontological reality: future events are as real as present or past ones, even though we have not yet experienced them. There is something to study about the future, something we can be, in principle, right or wrong about. On the other hand, precisely because the future is already ontologically real, the idea of alternative futures (or any such idea) becomes harder to sustain unless it is reinterpreted modally rather than temporally. We would have to say that we study possibilities, not the future itself.

This may sound like a natural adjustment, but notice the trap. When, let’s say, a foresight practitioner draws a branching diagram and labels the branches p1, p2, ... pn, the natural reading is that these are genuine alternatives, any of which could be realised depending on what happens between now and then. Eternalism does not allow that reading. There is no open set of branches waiting for one to be selected; one of p1, p2, ... pn is already there in the future, waiting to be reached rather than chosen. The other branches are not unrealised possibilities “competing” for actuality. They are simply ways things will not be. Whatever we are doing when we construct the branches, we are not mapping real alternatives in the way the practice usually presumes.

This forces a shift in vocabulary. We can still talk meaningfully about alternative futures, but the talk has to be about modality, about possibilities, as such, rather than about real temporal openness. The eternalist “block universe” contains one actual temporal sequence with the future already in the block, not a branching tree. The future is fixed in the block. This means that scenarios and other descriptions of the future become windows into a space of logical or epistemic possibility, not representations of real branches stemming from the present. This means, crucially, that their relationship to the single actually real timeline must be carefully specified. When Herman Kahn urged planners to “think about the unthinkable” (Kahn, 1962), he was, on the eternalist view, inviting them to explore a space of possibilities rather than literal temporal branches. This is not fatal to foresight, but it does mean the field needs an explicit account of what “possible” (and other such terms) mean.

The growing block: reality accumulates

The third theory related to time provides kind of a middle path. The growing block theory, developed originally by C. D. Broad (1923) and defended more recently by Michael Tooley (1997), says that the past and the present exist as parts of reality but the future does not. Reality grows: new events are added at a moving boundary fixed, at any given moment, by the present. The past is fully real and fixed; the future is genuinely open. In a sense, the futures cone, a traditional tool to convey futures studies ideas, embeds the growing block idea. The narrow end, the settled past, is a single trunk of accumulated reality. The present is the boundary point where the cone starts to open; the present is the branching point. And the widening cone represents possibilities that may be added to the fixed part of the reality in the future.

The growing block is intuitive in that it fits quite naturally the temporal experience that eternalism and presentism struggle with: we feel that there is a difference between the fixed or settled past and the open future. Take, for example, a foresight practitioner who has lived through an unexpected realisation of an unlikely scenario. That difference between a scenario, even as clearly formulated as possible, and it, is not merely conceptual but is something foresight responds to with the gut-feeling “that is why we formulated the set of scenarios, the future is not predetermined”. Of the three positions, it may appear that the growing block is the most natural philosophical home for futures studies, since it (i) preserves the openness that makes futures studies meaningful, while (ii) granting the past enough reality to serve as a stable foundation for analysis.

However, the growing block inherits the same truth-maker problem we saw with presentism: since no future this or that exists on this view either, it remains unclear what future-directed talk refers to, and what could make such a statement true – or better justified than another. Moreover, there must be some limits to how the growth happens – how do we figure these out? Some continuity with the past is to be taken for granted, but how much?

The choice among these three ontologies of reality of the point in time is not settled in philosophy, and it would be presumptuous for futures studies to tell what the answer is. However, futures studies can contribute here: How much we are able to know about the future and how are, in a sense, questions about futures studies as a field. And answering these questions tells us about the world surrounding us. A naturalist orientation in philosophy does not set grand questions a priori, but looks for the answers from fields of research – how they work and what makes, ontologically in this case, the work possible.

The choice, despite being difficult, is not neutral in its implications. Each position implies a different answer to what it means to “study the future”, and each shapes the kind of epistemic and practical claims a futurist can defensibly make – are the claims about the actual reality or abstract possibilities, for example. If they are abstract possibilities, why care? If they target the future that is “out there” but beyond our reach, why care – future is fixed and no scenario or other future-related work helps to shape it?

This goes back to very foundation of futures studies. For example, Bell’s nine core assumptions of futures studies – including the propositions that the future is not predetermined and that useful knowledge about it can be gained (Bell, 1997a) – carries an ontological claim about the future – but no real argument to back it up. It is most naturally read as a working assumption; but a completely wrong working assumption can be harmful, to say the least. A field that aims at intellectual maturity must at least recognise that its central concept, the future, is philosophically contested, and that the contest has practical implications for (i) how works in the field are framed, (ii) how possibility is understood, and (iii) how the relationship between present action and future outcomes is understood.

Can the Future Be Open? The Sea-Battle Problem

If the ontology of time determines whether the future exists as a part of the furniture of reality as it is, the problem of future contingents concerns the relationship between our ideas of openness and knowledge. And the relationship is one of the most consequential ones for foresight practice. Consider Amara’s three claims – that the future is not predetermined, not predictable, and can be shaped by choices. These are not self-evident truths. They are hard-won philosophical positions (if they are won), and the struggle to establish them is more than two thousand years old. The jury is still out about the winner.

The trouble starts with a deceptively simple question posed by Aristotle in his work On Interpretation (Chapter IX). Consider the pair of statements “Tomorrow there will be a sea-battle” and “Tomorrow there will not be a sea-battle.” Is one of these true today and the other false? If so, then the outcome is already settled: whichever statement is true today will be true tomorrow, and there is nothing anyone can do about it. It seems that the future is fixed. If one of the two statements has to be true now, then only one of those futures was ever really on the table; the other was never a possibility at all, just something we mistakenly imagined (see Øhrstrøm & Hasle, 1995). And the attempt to study alternative futures is conceptually confused from the start – or, at least, descriptions of alternative futures are mere fantasies – if this is the case.

In the generation following Aristotle, the problem was sharpened dramatically by Diodorus Cronus, who formulated the so-called Master Argument. Diodorus reasoned, roughly, as follows: Every true proposition about the past is necessary (what has happened cannot be undone). When the moment t comes and event E happens, not its alternative E, E becomes part of the past – E is necessary. Therefore, if we say that something is possible but never actually happens, we face a contradiction because the opposite of that something is necessary (meaning its opposite is impossible). To avoid the contradiction, Diodorus concluded that the possible just is that which either is or will be true. (See Gaskin, 1995). For futures studies, the Master Argument is a nightmare. If only what actually happens was ever possible, then every unrealised scenario, to mention just one famous example, in Jim Dator’s (2009) four archetypes – every Collapse that did not happen, every Transformation that remained imaginary – was not merely unrealised but strictly impossible. Possible futures would be a misleading label for something that was, in fact, impossible.

The discussion came more alive again in the Middle Ages, where logicians connected the problem to Christian theology. If God possesses infallible foreknowledge of future “free” choices, then a now-unpreventable truth about each choice appears already settled. This, the argument continues, obviously threatens human freedom and, thereby, moral responsibility. Determinism follows, or so it seems. Richard Lavenham, writing around 1380, gave a useful overview of the available positions that philosophers at that time saw (see more in Øhrstrøm, 1983). Lavenham classified the positions into four possible answers to the deterministic argument:

1. Accepting determinism outright – there is only one future waiting for us in reality.

2. Denying that future-tensed truths have always obtained. On this view, a statement like “there will be a sea-battle tomorrow” may be true or false today, but it was not true or false in the distant past before the battle. Truth values for future-tensed claims come into being over time rather than holding from eternity.

3. Rejecting the principle of excluded middle when it comes to the future – the idea that, for any future-tensed statement, either the statement or its denial is already true. Some statements about the future are neither true nor false right now, because reality has not yet settled the matter.

Think of the difference between options 2 and 3 as follows: Imagine a book being written. The story will end one way or another, but right now we ask: what is the status of the final chapter?

Option 2 says: the final chapter is already fully written – we just cannot read it yet. But here is the twist: the book itself only came into existence recently. So, while the ending is fixed now, it’s not as though the ending was fixed a million years ago, because back then there was no book. On the other hand, option 3 says: the final chapter is genuinely unwritten. It is not hiding from us; it does not exist yet. There is no fact about how the book ends, because the author has not decided. Only when the actual writing happens will there be something true to say about it.

4. Denying the necessity of the past – denying the intuitive idea that what has happened is now fixed and cannot be affected. This is the boldest move of the four, or so it seems. Why?

Writing before Lavenham, William of Ockham had already chosen option 4. Reasons were deep. As a Christian theologian, he was committed to divine foreknowledge: God knows from eternity everything that will happen. But knowledge requires truth, so if God knows today that there will be a sea-battle tomorrow, the statement must already be true today. Future contingents, therefore, must possess determinate truth-values. However, Ockham was equally unwilling to concede that the future is fixed, since human freedom demanded that the battle happening tomorrow be genuinely dependent on the people involved. Something in the deterministic argument had to give – and since foreknowledge blocked him from denying the truth of future-tensed statements, he denied instead the necessity of the past. His move here was clever. Not every past-tense statement, he argued, is really about the past. “Caesar crossed the Rubicon” genuinely is, and nothing can change it. But “yesterday it was true that there will be a sea-battle tomorrow” only looks past-tensed. It is not. Its truth really depends on what happens tomorrow. The principle that the past is fixed, according to Ockham, applies only to real past statements. Statements that disguise as being about the past but still describe the future remain open, because the future they describe remains open. But this raises a further question: what makes statements about the future true or false in the first place? To answer it, Ockham appealed to the notion of a true future: among all the courses of events that remain genuinely possible, one of them is the course that will actually unfold. That “singled-out” future is what makes present statements about tomorrow true or false, and what God eternally knows (See William of Ockham, 1983).

Seven centuries later, the debate is still alive. To mention one recent proposal that does something Ockham and his contemporaries never considered, we may learn from Barnes and Cameron (2009) who keep the future open without taking any of Lavenham’s four options. What they do instead of accepting one of the options is subtler: they locate the openness somewhere else entirely.

Take the statement “there will be a sea-battle tomorrow.” On the view of Barnes and Cameron, that statement is already either true or false right now – there is no third option, no gap, no missing value. But which of the two it is has not been settled yet. The world itself has not decided. The truth-value is there; what is missing is any fact about which value it has.

To understand this, let us go back to the example of the unfinished book. Option 2 said the final chapter was already written, merely out of our sight. Option 3 said the chapter was not written at all. Barnes and Cameron offer a third possibility, stranger than either. Imagine knowing, with certainty, that exactly one sentence is printed on the last page – either “they all survived” or “they all perished” – but there being, as yet, no fact about which of the two it is. The page is not blank. The ending is not hidden. It is unsettled, in the world itself, which of the two endings is the real one.

The payoff is that three claims that usually travel together can be unbundled. People defending an open future have typically also had to commit to one of three additional things: that logic somehow works differently when we talk about the future, that future events and objects do not yet exist in any sense, or that the laws of nature are not fully deterministic. Barnes and Cameron argue we need none of these assumptions. The future can be genuinely open even if standard logic applies, even if the laws are deterministic, and independently of whether future things “really” exist or not. The openness does not live in our logic or in our inventory of what exists. It lives in the world itself, so to say.

For futures studies, the moral is practical. Consider again how Dator’s (2009) unrealised Collapses and Transformations could have been genuine possibilities, really open futures, without our having to win any of the usual philosophical arguments first. We do not have to take sides on the metaphysics of time, or on the interpretation of physics, or on whether classical logic holds. We only have to take seriously a simpler and older thought: that reality can be, on some questions, genuinely unsettled – and that this unsettledness is what possibility, in the sense that matters for futures studies, actually amounts to.

The main consequences we wish to draw for foresight practitioners are twofold. First, the open future – the assumption that underwrites the entire field – is philosophically available, but it is not a free lunch. Every route to openness exacts a philosophical price, whether in logic, semantics, or metaphysics. Second, different ways of paying that price imply different formal architectures for the branching structure of alternative futures, and those differences matter for what can be said, consistently and rigorously, about the relationship between present choices and future outcomes. For example, we can say that we are able to choose a future; but this is different from saying the future is what it is because we choose in this or that way. The first requires openness, the second does not, while still giving us agency in the sense that the future is fixed all along through our actions.

Branching Futures: From Whiteboards to Formal Frameworks

Almost every person doing futures studies research has stood in a workshop and drawn a branching diagram on a whiteboard – a single line of shared history splitting into several paths in several directions – you have been doing philosophy, whether you knew it or not. There is a family of formal frameworks that is connected to this type of futures, called branching-time semantics, and gives rigor to our intuitive picture about branching futures. Its development is one of the quiet intellectual achievements of twentieth-century logic.

The idea can be tracked to the year 1958, when Saul Kripke, while being 17, would develop and provide the rigorous formal semantics for modal logic (i.e., formal rules of reasoning concerning propositions about possibilities, necessities, impossibilities, and so on). Given the philosophical state of mind at the time, this made the entire apparatus of necessity and possibility philosophically respectable – metaphysics was not highly regarded at the time. (Kripke, 1959, 1963) Kripke, in fact, wrote to A. N. Prior (whom we met above) and suggested that, in an indeterministic system, time should be represented not as a line but as a tree. A branching-time model pictures time as a tree. Each point in time – a “moment” – sits somewhere on this tree, and the complete paths running all the way through it are called “chronicles”. Each chronicle represents a full possible history of the world. The tree has a specific shape. Every moment has a unique past (if two moments precede a given moment, they must be comparable). The future, by contrast, branches: many possible continuations spread out ahead from the present. (See more detailed treatment in Øhrstrøm & Hasle, 1995; van Benthem, 1991).

The parallel to foresight practice, with multiple scenarios, paths, narratives, or what have you, is rather close as one may notice. Chronicles are, in a sense, the formal equivalent of fully developed scenarios, if we consider a scenario simply as “description of a future situation and the course of events which allows one to move forward from the actual to the future situation” (Amer et al., 2013, 23). The tree as a whole is the scenario space, i.e., scenarios taken in one context into account. Moreover, the linear ordering of events of the past that the formal system carries captures the intuition we already have mentioned: that the past is shared and settled, while the future is open and branches out into genuinely different possibilities.

Now, the question is how to handle truth-values within this branching structure; we have already seen the trouble the future truth-values, like in the case of the sea-battle, created. Several approaches in branching time semantics have been developed, each with, again, different philosophical costs and benefits. Consider, for example, Richmond Thomason’s “supervaluationist approach”. It offers one of the most elegant solutions. The idea is this: evaluate a future-directed statement not against a single timeline but against all the chronicles that pass through the current moment. If the statement is true on every chronicle, it is true, full stop. If the statement is false on every chronicle, it is false. Crucially, on this view, a future contingent may be neither true nor false at a given moment, even though a disjunction expressing excluded middle – such as “either there will be a sea battle tomorrow or there will not” – can still be true. The price of this result is the abandonment of ordinary truth-functionality (i.e., the truth of a compound statement is no longer determined solely by the truth-values of its parts, so a disjunction, the or, can be true even when neither disjunct is itself true). (Thomason, 1970) The great advantage is that this position preserves excluded middle when it comes to futures: the disjunction “either there will or there will not be a sea-battle” is true on every chronicle (on each one, one disjunct holds), even though, as we saw, neither disjunct is individually settled. We can say “several futures are genuinely live” without paradox. The cost in terms of the sacrifice of truth-functionality, the philosophical cost, may feel natural to those who practice foresight, while for philosophers it is harder to swallow. The lesson, however, is not in the details. The point we are driving at is that philosophy of futures studies must dig deep in order to establish with utmost rigor what the field is committed to.

Worthy of mention is that Nuel Belnap and his collaborators developed (2001) a closely related framework, known as STIT theory, short for “seeing to it that” – which extends branching-time semantics by incorporating agents and choices. In this, the theory is worthy reading for futures studies scholars and practitioners. On this picture, the future is represented as a tree of branching histories. The key distinction is between plain truth and settled truth: a claim is plainly true if it is true in one particular possible future branching out from the present, and settled true only if it is true in every possible future branching out from the present. For example, when Shell’s scenario team warned in 1972 that oil supply could be disrupted, they were not presenting the “could” outcome as settled, of course. They were identifying a possibility that was true in some histories but did not exclude others in which no such disruption occurred. This shows how close a formal framework can come to the study of the future: We can say without formal vacuum propositions such as “In this scenario, X happens,” without implying that X belongs to the one fixed future.

The framework becomes especially useful for futures studies once agency enters the picture. In STIT models, an agent’s choices at a moment split the possible futures into groups. Each group contains the histories that remain possible if the agent picks that option. On one particular STIT reading, i.e., in the deliberative STIT reading, an agent sees to it that something happens only if a choice of the agent guarantees the outcome (and the outcome was not already settled independently of that choice). This captures a distinction that is dear to futures studies: There are developments that decision-making of an agent can genuinely influence and those it must simply adapt to. In that sense, scenarios function like histories, while choice points mark the places at which agency can make a difference. However, the philosophical price of the system is that it can be secured only by sacrificing some intuitive logical or metaphysical commitment, like bivalence or ordinary truth-functionality.

Now, it does not matter for our purposes in this text which of these solutions is “correct” but what the debate as a whole reveals. The notion of an open future is available but not for free. It requires giving up at least one intuitively plausible logical or metaphysical principle; formal systems cannot leave gaps in them, so price tags are to be found. Different ways of securing an open future can be found in different formal architectures (only a few of which we even have space to mention here) for the branching structure of alternative futures. However, the informal picture – branches, scenarios, alternative pathways – has a rigorous formal counterpart. “Multiple futures” is not merely a metaphor to make workshops roll. It can reflect a formally representable structure with precise semantic properties, and understanding those properties is a necessary step to understanding the ultimate foundations of futures studies.

How Wide Is the Space of Possibility?

The previous sections asked whether the future is open. This section asks a different but equally important and difficult question: How wide is the space of possibilities? Even if the future genuinely branches, a further question remains about the nature, forming, and counting of the branches. When Bertrand de Jouvenel coined the term “futuribles” (“futurs possibles”) he literally stated that there are possibilities regarding futures and that futures studies focuses on them (de Jouvenel, 1967). But what kind of space is it? How broad is it? What are its boundaries? In philosophy, these questions are asked when possible worlds are discussed. There are three interesting theories about possible worlds we may discuss, and the choice among them has practical consequences for the scope of foresight.

The widest and most ontologically ambitious conception belongs to David Lewis’s concrete modal realism. Lewis argued that every logically consistent way things could be is a full, concrete physical universe. These universes are as real as our own but causally and spatiotemporally isolated from it. “Actual”, on his view, is merely an indexical when it comes to possibilities (like “here” is indexical). It picks out the world the speaker happens to inhabit, with no metaphysical privilege over non-actual worlds (Lewis, 1986).

Lewis proposes a principle of recombination to help secure modal plenitude: Roughly, any objects from any worlds can coexist in a single world via duplicates – you and I have duplicates in the possible world, and so does everything else. Then these duplicates can be combined in many ways, covering all possibilities. Critics argue that this recombination, since it proceeds by rearranging duplicates of what already exists, cannot by itself generate genuinely novel possibilities and thus cannot guarantee full modal plenitude. What about alien properties, for example, not to be found in our universe? For futures studies, Lewisian realism, while sounding odd, philosophically speaking, provides the maximal back-up. If every possibility is real somewhere, then the alternative futures studied by futurists are not just imagined – they actually exist as real worlds among countless others. The obvious objection is what Lewis himself called the “incredulous stare”: the sheer implausibility of an ontology of infinitely many concrete universes. Lewis did bite the bullet and argued that the theoretical benefits – a genuinely reductive analysis of necessity, possibility, and counterfactuals – justify the cost, but most philosophers prefer more modest ontologies, to say the least.

Abstractionism is a fundamentally different theory about possible worlds. On Alvin Plantinga’s account they are complete reality-descriptions or ways things might have been, not concrete objects. Plantinga’s basic move is: possible worlds are “maximal” possible states of affairs of which something either is true or not, not concrete universes. The actual world is the unique state of affairs that does obtain (Plantinga, 1974). For Plantinga, it is possible that there could have been, let’s say, a speaking duck. This just means reality could have been that way. It does not mean there is another concrete world where the duck is real. In Robert Adams’ version, possible worlds are treated as maximally consistent sets of propositions (Adams, 1974). The trade-off is that abstractionism does not reduce modality: possible worlds are defined in irreducibly modal terms (a world is a state of affairs that possibly obtains). While worrisome, this is not fatal; the abstractionist framework does not eliminate the concept of possibility but provides a structured explication of what possibilities (and necessities, and so forth) are. For futures studies, abstractionism provides a philosophically respectable way to talk about alternative futures. They are descriptions of possible states of affairs, and these states are abstract – we do not have to commit to the view that claims about the future must be about something as concrete as the world surrounding us. For example, to continue a bit further, any scenario can be understood as a partial sketch of an abstract possible world without anyone needing to believe that the scenario describes something concrete.

David Armstrong’s combinatorialism represents a third approach (Armstrong, 1997). In this approach possible worlds are recombinations of the actual world’s basic objects and properties. Any actual simple objects (my red shirt) and any actual simple universals (like blue) may combine in any arrangement (meaning we can say “my shirt could have been blue”). The worlds are not subject to any restriction other than formal consistency. This keeps possibility tied to non-modal ingredients already present in reality.

But it is here that the approach faces a significant limitation: it struggles to make sense of genuinely novel things that do not exist. “Alien universals” that cannot be found anywhere in our universe cannot be talked about – while it seems that there clearly is more to possibilities than what the real world has as constituents (perhaps there could have been a stone that has a wholly new kind of glow, call it “glitterenhanced”). For a field that often values opening up imagination and transformative change, this constraint is worth taking seriously. If it turns out that combinatorialism is the correct position, futures studies might be a castle built on sand. This is only mentioned to show why obscure philosophical issues must be taken seriously, if futures studies is to stand on as stable ground.

The contrast among these three approaches maps directly onto a practical question about the scope of foresight. If we accept Lewisian concretism, the imagination of futures studies is constrained only by logical consistency, anything goes (as long as it is not self-contradictory). If combinatorialism is chosen, alternatives must be built from ingredients already present in reality, and genuinely alien futures are harder to accommodate. In this regard, abstractionism provides a middle position that permits a wide set of possibilities without murky ontological commitments. Be that as it may, every time we, in futures studies, judge a scenario “possible” or “impossible”, an implicit modal ontology is at work. Ontological choices matter for what can be said, rigorously, about the future. We must be careful in this regard. Again, futures studies’ ways of working come with philosophical price tags.

How Do We Know What Is Possible?

Thus far we have discussed mainly the ontology and semantics of time, future, and possibilities. Some words about modal knowledge (i.e., knowledge about possibilities) need to be said after all that was written above.

Assume – for the sake of argument; we saw nothing comes for free – that the space of possibility is ontologically well-grounded and that the future is something we can make claims about. Futures studies still faces an epistemological challenge: how can we have modal knowledge? This question is pressing for futures studies because the field attempts to generate and assess claims about what is possible, plausible, or impossible. Not only is this how we speak, but what we do: when a Delphi panel evaluates the likelihood of a technological breakthrough, when a horizon scanner flags a weak signal as a harbinger of change (see e.g., Linstone & Turoff, 1975; van Veen & Ortt, 2021; Ansoff, 1975; Hiltunen, 2008), when a scenario team judges an outcome plausible or dismisses it as “too sci-fi”, modal claims are being made. What justifies such claims? Possibilities are not usually living in our world as such. How do we gain access to them?

One influential tradition grounds modal knowledge in conceivability. Stephen Yablo argues that conceivability provides (defeasible, can be overridden later) evidence for possibilities: Imagining of p is having the experience of p seeming possible – much as perceiving a tree is having the experience of a tree seeming to be there (Yablo, 1993). The idea is that imagination provides an appearance of possibility in the same way as perception provides an appearance of fact: fallible, but still evidential. In terms of futures studies, imagined alternatives are not self-validating, but they are not worthless either. The fact that a scenario can be coherently imagined counts epistemically toward it being possible, even though further analysis is needed to determine whether it describes a real possibility.

Timothy Williamson has developed another influential alternative that grounds modal knowledge not in conceivability but in our capacity for counterfactual reasoning. On his account, we can know what is possible by imaginatively developing “what if?” suppositions against background knowledge. We create a counterfactual scenario, insert into that scenario what follows, and see what follows. “Had I put on a blue shirt, I could have worn a red shirt in the afternoon” makes “wearing a red shirt” possible – I could have changed my shirt, for all we know. This capacity to counterfactual reasoning, Williamson argues, is deeply integrated into almost all central cognition, most of which is directly relevant to futures studies. Planning, decision-making, and scientific theorising all involve asking what would happen if things were different, i.e., counterfactual questions (T. Williamson, 2007). For futures studies, this is one of the strongest philosophical bridges from abstract metaphysics to practical foresight. The cognitive capacity practitioners exercise when building scenarios is, on Williamson’s account, the same capacity that gives us access to the modal realm.

Interestingly, this shows how philosophy is relevant to more specific ideas in futures studies than merely to shared modal struggles. Notice how Riel Miller’s concept of “futures literacy”, which refers to the cultivable ability to use the future effectively in the present (R. Miller, 2018) – can be grounded philosophically, given the discussion above. Futures literacy is not merely a professional skill but the development of a fundamental human capacity for understanding and using possibilities in their lives broadly.

The philosophical approach on modalities also provides a formal treatment for counterfactuals, the “what-if” analysis. Robert Stalnaker (1968) and David Lewis (1973) independently developed the so-called similarity analysis of counterfactuals: a counterfactual is true when the closest possible worlds in which the antecedent holds (“had it been the case…”) are worlds in which the consequent (“it would have been the case”) also holds. To answer ‘what would happen if oil prices doubled?’, one needs to focus on the possible world most similar to ours in which oil prices double, and ask what follows. Good counterfactual reasoning thus requires reliable background knowledge and a sound sense of which possibilities are “closest”. Defining closeness has proven difficult, but we do not go into that now. What matters to us is how well someone can recognise when a counterfactual world and its workings are closest to us – and one could argue that this is why experienced people in futures studies, well acquainted with the domain-related knowledge, produce better scenarios than novices. Opening up imagination is one thing; following rigorously “what-if” paths is completely another thing (Virmajoki, 2024).

Anand Vaidya and Michael Wallner (2021) usefully distinguish between the Access question – how we initially gain epistemic access to modal claims – and the Navigation question – how we move from knowledge of one kind of possibility to another. They also identify what they call the problem of “modal epistemic friction”: what constrains the movement from imagined possibilities to stronger claims about what is really possible? This distinction matters in futures studies. Consider a situation where a scenario team brainstorms. They are accessing possibilities – opening a space of alternatives. When they evaluate those alternatives for, let’s say, plausibility, they are navigating between possibilities, moving from broad possibilities toward claims about plausibilities that are more “realistic”. And the friction problem then sharpens the difficulty further: What allows (makes it non-arbitrary) the team’s leap from “I can imagine this” to “this is genuinely possible”? This is, in a sense, the quality-control question for all claims about future possibilities. Peter Schoemaker (1995) argued that the value of scenarios lies in disciplining the imagination; easier said than done, would a philosopher say.

Not everyone is optimistic about the reach of our modal capacities. Peter van Inwagen (1998) argues, by analogy with the limits of vision, that we should be skeptical about modal claims beyond the reliable range of our cognitive abilities. Undeniably, the argument goes, we can trust everyday modal judgments – that I could wear a red shirt, that the game is cancelled – but we should withhold confidence about what could be called exotic or far-from-actual (“far-fetched”) possibilities. If Van Inwagen is right, the field’s ability to explore genuinely novel futures is badly damaged – at least we need to stay away from saying that we have any claim for knowledge about genuinely novel futures.

What is said above are general considerations about modal knowledge. As such, they tell us that philosophical work is done when it comes to existence of the future and possibilities and our knowledge of them, but philosophical positions on the matters are not straightforward. To build futures studies on ideas like “the future is open”, “we can affect the future”, “we study possible futures”, the philosophical underpinnings of such claims must be made clear rather than taken for granted. Philosophy is not a bedrock, however. It has consequences. Before wrapping up, we wish to sketch some of the consequences – how to move from abstract and sometimes nitpicking-sounding philosophy to research practices in futures studies.

From Philosophy to Practice

The philosophical issues discussed bear surprisingly directly on the intellectual coherence and practical credibility of futures studies as a field, as we have pointed out throughout. Three further but connected ideas about the bridge are to be emphasised.

First, while we have not discussed the notion of understanding as such, a common idea is that understanding requires the ability to deal with possibilities. For example, causality is often understood as a dependency-relation concerning what would have happened to the effect if the cause had been different (e.g., Woodward, 2003). In general, the idea that understanding can be defined as the ability to navigate spaces of possibility provides one suitable philosophy for futures studies. According to Soazig Le Bihan (2017), a person has (modal) understanding of a phenomenon when that person possesses the know-how to move around within some region of the space of possibility associated with that phenomenon. We do not have understanding when we can see possibilities here and there, but we have to competently navigate a space of possibilities. Consider a traffic jam: knowing one way it can arise (for example, a rush-hour collision) provides some modal understanding, but robust understanding is associated with navigating a wider space, for example, understanding what would happen if the driving lines were rearranged or if the collision can be solved quickly.

This, in a sense, captures much of what futures studies, at least, aims to do. Consider the following words that are worthy to cite at length:

“Long-term planning often demands an understanding of multidisciplinary connections, for example the interdependence of demographic, social, technological, economic and political developments. [--] However, more is needed for strategic decisions. They demand an overall image that expresses in which combinations the lower, middle or upper variants of the disciplinary prognoses can possibly occur, because it is clear that for such quantities as have been mentioned above the variants do not occur or fail to occur independently of each other”. (Weimer-Jehle, 2006, p. 335 [emphases added]).

The idea is similar: to provide understanding relevant to decision-making through futures studies, we need to navigate spaces of possibility. Moreover, morphological analysis (e.g., Ritchey, 2018), or system dynamics (e.g., Sterman, 2000), to mention some approaches, are not merely listing possibilities. They aim at understanding what we are navigating and how. Philosophy provides tools to understand how and why this “modal navigation” happens; or how it can happen in the first place. Grasping the modal structure of the system under study: which configurations are possible, how they depend on one another, and where the leverage points lie. Futures studies, on this view, is not an exotic supplement to ordinary inquiry but a natural expression of the drive toward understanding.

Second, James Woodward’s interventionist account of explanation treats explanations as answers to “what-if-things-had-been-different” questions (Woodward, 2003). This way of thinking can relatively easily be turned into future-oriented thinking: we ask about a system, what would happen, if this very particular variable of the system was changed. Consider a foresight team that asks how the world would change if a key assumption does not hold in the future; for example, if a technology matured a decade earlier than expected, if a political coalition collapsed, if a pandemic struck. Here, they are, at best, engaging in the kind of reasoning that Woodward identifies as the core of good explanatory practice. The team generates explanations of possible states of affairs in the future. By providing understanding they provide useful future-knowledge. The study of alternative (or possible, or what have you) futures is not separate from scientific explanation but intellectually continuous with it.

Third, Sohail Inayatullah’s Causal Layered Analysis (Inayatullah, 1998) can be understood, in the light of the discussion in this text, as a relatively different type of method to navigate possibilities: It does so by grounding them, interpretatively speaking, on different levels of our ways to see and “read” the world at different depths. CLA’s four layers, i.e., litany, causal, discourse and worldview, and myth and metaphor, may be read, if we wish to interpret in the light of the philosophy in this text, as corresponding to progressively deeper levels at which the space of possibility is shaped and constrained. At the surface level, alternative futures are variations on current trends: rearrangements of familiar elements, close to what Armstrong’s combinatorialism says. At the level of causes, we may use interventionism to understand what happens in our thinking. At the level of worldview, we may need to go to the very basis of the questions, such as Williamson, among others, answers, about ways of grounding possibilities in the first place. At the level of myth and metaphor, we may need to question if our modal thinking is on the right tracks in general: whether we are restricted by current understanding of modalities. For example, genuinely alien possibilities that combinatorialism struggles to accommodate (see above) may be probed at on this level. If we do this and read CLA through the lens of modal epistemology, it becomes a method for navigating space of possibility at multiple levels of depth, and the philosophical discussion of conceivability, counterfactual reasoning, and modal friction are tools to make sense of what it means to navigate possible futures at different depths.

Together, these three points, that are an admittedly limited set of examples, point toward the possibility of bridging the ontological issues and foundations discussed in this text and futures studies methods. Ontological questioning is not an obstacle to the practice of futures studies but a resource for it. Understanding systems in terms such as spaces of possibility, cultivating the ability to reason counterfactually, and recognising that good explanation is inherently providing understanding about alternatives are important to see the field as seriously thought-through. The issues discussed in this text are not exotic philosophical doctrines or decorations but attempts to map what competent foresight practitioners already do. What philosophy adds is precision, self-awareness, and a set of tools for distinguishing well-grounded possibility claims from wishful thinking and far-fetched pondering.

Conclusion

This text has argued that the philosophical underpinnings of futures studies are neither trivial nor optional. The question of whether the future exists at all – answered differently by presentism, eternalism, and the growing block theory – is central to what it means to study something that may or may not (yet) be part of reality. Aristotle’s sea-battle problem, debated across centuries for the logical and theological implications of such problems, shows that the openness of the future is not to be taken as given but is a deep philosophical commitment. We must purchase it at the cost of giving up at least one intuitively plausible philosophical principle, it seems. That makes the philosophy of futures studies much more exciting: what are we willing to be to keep the field going, philosophically speaking?

Moreover, the metaphysics of possible worlds can be used to understand how wide the space of possibility (if any exists) the field can claim to engage with. Lewis’s concrete modal realism provides the maximally expansive backdrop; Plantinga’s abstractionism offers breadth without concrete ontological commitment; and Armstrong’s combinatorialism ties possibility to the rearrangement of actual ingredients. Then, associated with the metaphysics of possibilities, modal epistemology – from Yablo’s conceivability as defeasible evidence about possibilities through T. Williamson’s (2007) grounding of modal knowledge in counterfactual reasoning to the skeptical challenges of Van Inwagen and others – is badly needed to understand what allows and, equally importantly, constrains the movement from imagination to justified claims about real possibilities.

For futures studies, these are not abstract puzzles to be dismissed as being someone else’s problems. Every time a futures studies scholar or a foresight practitioner distinguishes, let’s say, a plausible scenario from an implausible one, or a possible future from an impossible one, philosophical commitments about time, truth, possibility, and knowledge are being made, whether we like it or not. Making the field’s ontological commitments visible and debatable; showing that they are not mere assumptions but positions with arguments behind them and alternatives beside them – is the contribution this text has sought to make. The lesson is, as long as we discuss possibilities and the future (as open), we make ontological commitments, whether we say them aloud or not. And to build the field’s ontological foundations, one needs to say something. After speaking aloud our position, we must put our money where our mouth is, and only do ontologically sound futures studies without tripping on one’s own feet.

Bibliography

Adams, R. M. (1974). Theories of actuality. Noûs, 8(3), 211–231.

Amara, R. (1981). The futures field: Searching for definitions and boundaries. The Futurist, 15(1), 25–29.

Amer, M., Daim, T. U., & Jetter, A. (2013). A review of scenario planning. Futures, 46, 23–40.

Andersson, J. (2020). Ghost in a Shell: The scenario tool and the world making of Royal Dutch Shell. Business History Review, 94(4), 729–751.

Ansoff, H. I. (1975). Managing strategic surprise by response to weak signals. California Management Review, 18(2), 21–33.

Armstrong, D. M. (1997). A world of states of affairs. Cambridge University Press.

Barnes, E., & Cameron, R. (2009). The open future: Bivalence, determinism and ontology. Philosophical Studies, 146(2), 291–309.

Bell, W. (1997a). Foundations of futures studies: Human science for a new era. Volume 1: History, purposes, and knowledge. Transaction Publishers.

Bell, W. (1997b). Foundations of futures studies: Human science for a new era. Volume 2: Values, objectivity, and the good society. Transaction Publishers.

Belnap, N., Perloff, M., & Xu, M. (2001). Facing the future: Agents and choices in our indeterminist world. Oxford University Press.

Broad, C. D. (1923). Scientific thought. Kegan Paul, Trench, Trubner & Co.

Cournand, A., & Lévy, M. (Eds.). (1973). Shaping the future: Gaston Berger and the concept of prospective. Gordon and Breach.

Dator, J. (2009). Alternative futures at the Manoa School. Journal of Futures Studies, 14(2), 1–18.

de Jouvenel, B. (1967). The art of conjecture (N. M. Lary, Trans.). Basic Books.

Gaskin, R. (1995). The sea battle and the master argument: Aristotle and Diodorus Cronus on the metaphysics of the future. De Gruyter.

Hancock, T., & Bezold, C. (1994). Possible futures, preferable futures. Healthcare Forum Journal, 37(2), 23–29.

Hiltunen, E. (2008). The future sign and its three dimensions. Futures, 40(3), 247–260.

Inayatullah, S. (1998). Causal layered analysis: Poststructuralism as method. Futures, 30(8), 815–829.

Kahn, H. (1962). Thinking about the unthinkable. Horizon Press.

Kripke, S. A. (1959). A completeness theorem in modal logic. Journal of Symbolic Logic, 24(1), 1–14.

Kripke, S. A. (1963). Semantical considerations on modal logic. Acta Philosophica Fennica, 16, 83–94.

Le Bihan, S. (2017). Enlightening falsehoods: A modal view of scientific understanding. In S. R. Grimm, C. Baumberger, & S. Ammon (Eds.), Explaining understanding: New perspectives from epistemology and philosophy of science (pp. 111–136). Routledge.

Lewis, D. (1973). Counterfactuals. Harvard University Press.

Lewis, D. (1986). On the plurality of worlds. Blackwell.

Linstone, H. A., & Turoff, M. (Eds.). (1975). The Delphi method: Techniques and applications. Addison-Wesley.

Markosian, N. (2004). A defense of presentism. In D. W. Zimmerman (Ed.), Oxford studies in metaphysics (Vol. 1, pp. 47–82). Oxford University Press.

Miller, R. (Ed.). (2018). Transforming the future: Anticipation in the 21st century. UNESCO Publishing/Routledge.

Øhrstrøm, P. (1983). Richard Lavenham on future contingents. Cahiers de l’Institut du Moyen-Âge Grec et Latin, 44, 180–186.

Øhrstrøm, P., & Hasle, P. (1995). Temporal logic: From ancient ideas to artificial intelligence. Kluwer.

Plantinga, A. (1974). The nature of necessity. Oxford University Press.

Polak, F. L. (1973). The image of the future (E. Boulding, Trans. & abridg.). Elsevier Scientific Publishing Company.

Prior, A. N. (1953). Three-valued logic and future contingents. The Philosophical Quarterly, 3(13), 317–326.

Prior, A. N. (1967). Past, present and future. Clarendon Press.

Putnam, H. (1967). Time and physical geometry. Journal of Philosophy, 64(8), 240–247.

Ritchey, T. (2018). General morphological analysis as a basic scientific modelling method. Technological Forecasting and Social Change, 126, 81–91.

Schoemaker, P. J. H. (1995). Scenario planning: A tool for strategic thinking. Sloan Management Review, 36(2), 25–40.

Sider, T. (2001). Four-dimensionalism: An ontology of persistence and time. Oxford University Press.

Stalnaker, R. C. (1968). A theory of conditionals. In N. Rescher (Ed.), Studies in logical theory (pp. 98–112). Blackwell.

Sterman, J. D. (2000). Business dynamics: Systems thinking and modeling for a complex world. Irwin/McGraw-Hill.

Thomason, R. H. (1970). Indeterminist time and truth-value gaps. Theoria, 36(3), 264–281.

Tooley, M. (1997). Time, tense, and causation. Oxford University Press.

Vaidya, A. J., & Wallner, M. (2021). The epistemology of modality and the problem of modal epistemic friction. Synthese, 198(Suppl. 8), 1909–1935.

van Benthem, J. (1991). The logic of time: A model-theoretic investigation into the varieties of temporal ontology and temporal discourse (2nd rev. ed.). Kluwer Academic Publishers.

van Inwagen, P. (1998). Modal epistemology. Philosophical Studies, 92(1–2), 67–84.

van Veen, B. L., & Ortt, J. R. (2021). Unifying weak signals definitions to improve construct understanding. Futures, 134, 102837.

Virmajoki, V. (2024). On the function and nature of historical counterfactuals: Clarifying confusions. Journal of the Philosophy of History, 19(1), 80–104

Voros, J. (2003). A generic foresight process framework. Foresight, 5(3), 10–21.

Wack, P. (1985a). Scenarios: Uncharted waters ahead. Harvard Business Review, 63(5), 72–89.

Wack, P. (1985b). Scenarios: Shooting the rapids. Harvard Business Review, 63(6), 139–150.

Weimer-Jehle, W. (2006). Cross-impact balances: A system-theoretical approach to cross-impact analysis. Technological Forecasting and Social Change, 73(4), 334–361.

William of Ockham. (1983). Predestination, God’s foreknowledge, and future contingents (M. M. Adams & N. Kretzmann, Trans.; 2nd ed.). Hackett.

Williamson, T. (2007). The philosophy of philosophy. Blackwell.

Woodward, J. (2003). Making things happen: A theory of causal explanation. Oxford University Press.

Yablo, S. (1993). Is conceivability a guide to possibility? Philosophy and Phenomenological Research, 53(1), 1–42.

Introduction

Assume, for the sake of argument, that a scenario team gathers around a whiteboard. Someone has sketched four futures for the energy system in 2050, and the facilitator asks the group to sort them: which of these are possible? Which plausible? Which probable? And which desirable?¹ Imagine the room’s opinion are split. One participant says that she thinks that total grid decarbonisation by 2050 is plausible, since no known physical law rules it out. Another objects, correctly it seems, that plausibility asks for more than the absence of impossibility – it requires we can identify a causal pathway, mechanism, that could actually get us there. A third participant cynically adds that the scenario is not even possible in any sense that matters, given current political realities; a fourth answers that political realities are precisely the kind of thing scenarios are met to challenge, so referring to political reality is not an argument against the scenario as such – desirability affects what can be made possible. Twenty minutes on, the team has learned a great deal about one another’s assumptions. However, they have learned nothing at all about what the words they use – plausible, possible, desirable – mean.

This is not a failure of facilitation, people learn. It still is a failure of vocabulary, and that does matter. Let us explain in this text.

The modal notions central to futures studies practices and communications – possible, plausible, probable, preferable – are at the same time (i) the field’s most indispensable concepts and (ii) its most dangerously imprecise. These notions are expressed in nearly every canonical framework. For example, Voros’s (2003) futures cone (which already exist at least in Hancock and Bezold’s 1994 work), sorts futures into nested zones – and these are zones of possibility, plausibility, probability, and preferability – maybe we can add desirability, and layer of preposterous. Amara’s (1981) three commitments – that the future is not predetermined, the future is not predictable, and future outcomes can be influenced by present choices – presuppose that one can draw meaningful distinctions between what is predetermined, what is predictable, and what is open to influence. Finally, Dator’s (2009) four archetype futures – Continued Growth, Collapse, Discipline, Transformation –sorts, at least implicitly, futures in terms of their relation to present trajectories and ask us to think type of alternatives – modally loaded term.

And in none of these frameworks are the terms given definitions sharp enough to settle the disagreement that opens this text. The words are living the life of their own in actual work in futures studies done every day but rarely anyone has said what the words mean with philosophical clarity.

Behind these terms lie metaphysical questions about the reality of the future, the openness of what is to come, and the extent of the space of possibility. These are questions about the furniture of reality – the metaphysical architecture that makes talk of alternative futures (a modal notion) coherent at all – and they are bracketed in what follows. This text asks a downstream question: given some space of possibilities, how should the vocabulary for moving in that space work? What does it mean, exactly, to call a future possible rather than plausible, or plausible rather than probable? What kind of reasoning is associated with the modal notions in the field?

Due to the various and even contradictory use of the modal notions in the field, we do not claim to track their usage to some ultimate point of departure or their “real meaning”. The tool we use in this text is what in philosophy is called conceptual engineering – the deliberate “sharpening” of concepts so that they are fit better to certain practices, such as thinking about futures. The goal is not legislation of meanings from outside. Rather, by connecting the use of a concept to its meaning through “engineering”, and asking what each term must mean if it is to do what futures studies does. We proceed term by term: discuss possible, then plausible, then probable, and touch on desirable. Along the way we run into a form of reasoning that underwrites much of the practice (or so we argue): counterfactual reasoning, the disciplined “what if things had been different?” thinking, which is can be associated with the engine of scenario construction.

Conceptual Engineering and the Futures Vocabulary

Before sharpening any single term, we must steer clear about the kind of work is to follow. Conceptual engineering, as it is practised in recent analytic philosophy, begins the observation concepts we “inherit” are not always fit for purposes we use them. They are vague where we need precision, ambiguous where we need clarity, or loaded with connotations that quietly affects research and its use (see e.g., Cappelen, 2018; Haslanger, 2012). The term conceptual engineering itself was, as is fitting to its own meaning, coined more than once – independently within “Carnap scholarship” (Carus, 2007; Creath, 1991) and within “metaphilosophy more broadly” (Blackburn, 1999; Floridi, 2011). The two traditions have since converged, at least somewhat, on a shared programme of kind of concept-improvement (Brun, 2016; Isaac, Koch, & Nefdt, 2022). Conceptual engineering does not to describe how a concept happens to be used – that is conceptual analysis, the older and more “descriptive” tradition – but to ask how it should be used, given what we need it for (Nado, 2021; Chalmers, 2025). Philosophy here becomes a “translational”, problem-solving thinking aimed at workable solutions to conceptual problems (Nado, 2021; Blackburn, 1999). The question shifts: not “what do people mean by X?” but “what should X mean, if it is to do perform certain functions?” – a move that treats concepts not as fixed objects of contemplation but as devices whose functional efficacy can, and should, be improved (Nado, 2021; Isaac et al., 2022; Woodward 2003).

The process has a roughly the following nature. At a minimum it runs through four phases: describing the target concept, assessing how it currently performs it function, improving where it fails or causes confusion, and implementing the revised version in the community (like futures studies scholars and foresight practicians) that uses it (Isaac et al., 2022; Cappelen & Plunkett, 2020). Drawing on Carnap’s (1950) “method of explication”, we may treat these phases not as forming a tidy sequence but, rather, we move back and forth between the phases until we achieve what we want (e.g., Brun, 2016; Chalmers, 2025; Isaac, Koch, and Nefdt, 2022; Thomasson, 2025).

One lesson for our purposes can be pinpointed for our purposes: the goals of futures studies as field and practice are plural. In conceptual engineering, one may wish to change how people classify things, or what inferences they draw, or the truth-conditions of key sentences, or the norms of usage – and each of these aims, it seems, requires a different method (Isaac et al., 2022; Koch, 2021; Nado, 2023). The framework of conceptual engineering is, however, “meta-normatively neutral”: it works with whatever substantive purpose one may have, but it insists the purpose has to be made explicit during the engineering process so it works (see e.g., Dutilh Novaes, 2020; Haslanger, 2005).

Futures studies is an exceptionally good candidate for the work of conceptual engineering. The field draws its modal vocabulary from several traditions, sources, and intuitions, Often uses of “possible” and “probable”, for example, differ from people to people, from project to project, from one futures studies tradition to another. Consider this: when a logician says something is “possible”, she means, more or less, that it is not self-contradictory. When a strategist in a company says a scenario is “possible”, she means something far richer – something along the lines that the scenario consistent with what is known about the world and is reachable by at least a rough causal (that can be at least sketched) pathway from here. When an ordinary person says “it is possible that tomorrow rains”, she may be expressing uncertainty (given weather forecasts), not making a claim about logical consistency or causal routes. Three concepts are thus travelling under one word – and we are in the danger of moving between them, more or less, without warning. This is the kind of situation conceptual engineering was tailored to diagnose and repair (Cappelen, 2018; Isaac et al., 2022).

The stakes are not merely academic, that we do consistent research reporting. When a foresight deliverable calls a future plausible to an audience – a board of directors, a government ministry, a community planning body – the stakeholders deserve to know what kind of meaning the word has in that case and what is the evidence and reasoning that are taken to warrant plausibility-claim. For example, if plausible would only mean “not obviously self-contradictory”, the warrant is thin. If it means “an identifiable chain of causes could lead from here to there”, the warrant is considerably heavier (assuming the chain is well-reasoned). If it means “some expert assigns it non-trivial probability”, it is heavier again but different in kind (no causal reasoning behind the claim but trust to experts). The current vocabulary is such that all of these uses are allowed. And here is where conceptual engineering is needed: such ambiguities are not innocent (Cappelen, 2018; Queloz & Bieber, 2022). When one word is chewed with three different epistemic standards in three mouths, the risk is not merely that people talk past one another. It is that decisions based on “plausibility” get murky and contradictory even.

There is an interesting discussion about the history of the most loaded word in the lot. Ramírez and Selin (2014), tracing the etymology of “plausibility” and “probability” in English, show that the two ran together until the seventeenth century and pulled apart “only” over the following three hundred years. “Probable”, once tied increasingly to mathematical and statistical reasoning and representations, gained a kind of scientific nobility. “Plausible” kept, from the start, an association with mere appearance. This can be understood once we notice that it is a cousin to applause, act which indicates the winning of public approval. In its earliest English uses it carried the hint of a “false appearance of veracity”, and – no matter how ironical it may sound in our current world – this a lineage that ends, more or less, in the Cold-War coinage plausible deniability: a polite name for institutional lying.

This means that a term the field of futures studies has at its core, plausibility, and associates with certain presentations of future, is loaded with certain associations – looking good rather than being true, winning people over, sounding convincing while being empty. The word, one might say, has a odd past when mirrored to its current use in futures studies. And that is precisely the type of a mismatch between inherited meaning and required function that conceptual engineering was built to repair.

In what follows, we ask: (i) what work does this or that concept do, or what function it has, in futures studies, and (ii) how could we provide a meaning to the concept that allows to do that work or satisfy the function? How we could and should use a concept will not likely correspond its current meaning-bundle, that is the point, but rather we want to improve the rigour, the clarity, and the decision-relevance of the concept. We begin discussion concerning the concepts by very rough mapping of how the term is used across the field’s representatives (description of usage comes first, as Isaac et al. 2022 argue) before saying how it should be perhaps be refined. And we keep one ethical eye open throughout: conceptual engineering is tied into conceptual ethics (Burgess & Plunkett, 2020). Conceptual engineering, as M. Shields (2021) and Cappelen (2018) warn, can become akin to conceptual domination if its ends are hidden and its proposals are shielded from criticism from the very people they are meant to serve. This means that the proposals we provide are offered as an ongoing, participatory conversation, not handed down as orders.

“Possible”: More Than Merely Conceivable

Of all the modal terms, possible may appear as the least controversial. It is not. The metaphysics of possibility includes Lewis’s (1986) concrete modal realism, Plantinga’s (1974) abstractionism, and Armstrong’s (1997) combinatorialism. Lewis treats possible worlds as concrete realities, Plantinga as maximal possible states of affairs, and Armstrong as recombinations of actual objects and properties. Each provides a different account of what possible worlds are and how vast is the space of possibility. These accounts are ontological: about what exists. The question now is conceptual: about what practitioners mean, and should (could) mean, when they stamp a future “possible”.

In philosophical usage, several well-separated varieties of possibility have been distinguished, and the differences matter enormously for futures studies – this is perhaps the most crucial place where philosophy should be studied by those who work on futures. Modal epistemology fixes mostly on the so-called alethic or “objective” modalities – possibility is grounded on of how the world is, not of how it is believed to be. Among the principal three alethic possibilities are logical, physical, and metaphysical possibility. (Roughly: logical = not self-contradictory conceptually; metaphysical = how reality could actually have been deep down; physical = allowed by the laws of nature).

The “textbook picture” is to nest these: physical possibility inside metaphysical, metaphysical inside logical (Kripke, 1980; Hale, 2013). The boundaries, it must be noted, are not settled. According to Chalmers (2002) metaphysical and logical (conceptual) modality coincide; according to Shoemaker (1998) metaphysical modality deflates so that it is co-extensive with physical modality; and Priest (2021) questions whether there is a notion of metaphysical necessity that is distinct from conceptual necessity or physical necessity. While the picture can thus be contested, we can use it anyway.

For futures studies the practical question is a blunt one: which variety of possibility draws the outer boundary of the space of possibilities we must deal with?

If the answer is logical possibility, the space is vast, too vast, – anything that does not contradict itself is on the table. Escaping to metaphysical possibilities does not help much. Even those futures that break every known law of physics are included. A world in which people can fly are metaphysically possible. Obviously, these are mostly too wide to be much use. However, issue is not that simple. These logical possibilities that are far-fetched may earn their place in the “most speculative” exercises, for example, at the mythic floor of Inayatullah’s (1998) Causal Layered Analysis; where myth and metaphor probe the very edge of what we can think at all. To call these “speculative” is not criticism, rather, we wish to point out that sometimes even logical possibilities form the vast space we must take into account. It is not easy, but nothing valuable is easy, some would say.

Then again, if the answer is physical possibility, the space is narrower but still enormously vast: the laws of physics license a staggering range of social, political, technological, and ecological arrangements (for example, people have no names but are all identified by number, to give a toy-example). The answer must, then, be some sort of a “practical possibility” – consistent not only with natural law but with the world as it actually now stands and work. The space narrows again, to the futures one could actually reach from where we stand. It is this last boundary that everyday futures studies more or less lives inside of. This is, then, what futures studies scholars and people thinking about mostly mean when they talk about possible.

Practical possibility is a peculiar thing, and a few philosophers approach it almost head-on. The argument is that we grasp what is possible not through elaborate reasoning but more or less directly. One strand of this “head-on” approach is perceptual. Strohminger (2015) argues that perception itself can provide knowledge of possibilities; seeing a low branch we in some sense see that it could be climbed. A second strand is agentive rather than perceptual. Vetter (2023) argues that our sense of what is possible is rooted in the experience of our own action: we know what we can do, and the wider idea of what is possible grows out from this experience.

Of course, philosophers try to make sense of possibilities, like practical possibilities going to their roots – nobody would say (credibly) that they can see the future or feel how they change it; that would be plain silly. However, the arguments may be more relevant to foresight than they sound. When one tracks weak-signal or does horizon-scanning, one is trying to peak or have some feel what might be going on now and, thereby, in the future. When Ansoff (1975) brought weak signals into strategic management, he was pointing at something we might call “a perception of practical possibility”: an early indication that a development is possible, before anyone can say whether it is plausible or probable – these are separate issues. Hiltunen (2008), referring to Peirce, can be seen as sharpening the point: a future sign (distinguished from weak signal due to the inflation of meanings of the latter) consist of three parts – a signal, an issue, an interpretation. This would mean that the “perception” of practical possibility is interpretation-laden from the start. This, in itself is not a problem: even scientific observations are theory-laden. So there are ideas where practical possibilities are “seen” and form a seed for understanding futures.

A more generally used route to knowledge about possibility runs through the epistemology of conceivability. In brief, Yablo (1993) took what we can coherently imagine to count in favour of – though never to settle – what is possible: imaginability is evidence, but evidence that can be overturned. This is, near enough, what a brainstorming session does; it uses imagination as an access way to understand what is possible. But the route is not safe, to be sure, and we know that. The move from “I can imagine this” to “this is possible” is problematic, at least when we talk about practical possibilities. But the distinction has even philosophical problems. Vaidya and Wallner (2021) discuss “modal epistemic friction”. There must be constraints that keep the leap from “imagine” to possible” away from being arbitrary. Something can seem conceivable and turn out, on a closer look, to be impossible. One can imagine me having different parents but I could not – different genes, different person. And something can fail to seem conceivable to a particular group and be perfectly possible all the same; this would be a failure of imagination, not a fact about the world – we cannot conceive a science that is not there – describing it would be, paradoxically, to have it – but it is possible that science changes. Van Inwagen (1998), in short, compares modal judgment to seeing: at the far edge of the visual field we withhold verdicts, and the same withholding is due once modal claims stretch past what our minds handle with any reliability. In terms of future: The more exotic the future, the less our being able to picture it tells us. We can picture, imagine, freely. The world is under no obligation to comply or follow.

What this means for practice is clear enough. Calling a future “possible” is not just something one can say when nothing else can be argued for. It is an epistemic claim – that, at minimum, so far as we know, nothing rules this future out – and its epistemic weight (how seriously we take it in decision-making) depends entirely on the variety of possibility in play – logical, metaphysical, physical, “practical”. A future that is “logically possible” carries less decision-relevant weight than one a future that is “physically possible”, which carries less than a future that is “practically achievable from where we stand” (how we ever can know this; in this text we talk about vocabulary). One must be case-by-case engineer here: when one speaks about “possible”, one has to tell what one means by this and be consistent. If one is shown one’s vision for “practically possible” future is not really practically possible, one cannot back down and say that the future, at least, non-contradictory, one commits intellectual crime. The future may be lifted to a table in another context where brainstorming is more welcomed, but that does not mean mistake is erased. Possible is tricky, but it can somewhat be tamed through consistency.

“Plausible”: The Case for Causal Pathways

If possible is, when taken in the widest sense, the outer boundary of what cannot be ruled out, plausible is meant to form a more interesting zone of futures: the futures that earn serious analytical attention. In Voros’s (2003) futures cone the plausible sits inside the possible and outside the probable. In practice, it is perhaps the most used but least consistently defined word in the field; in a sense, futures studies is the study of plausible (not just trend extrapolation, no mere speculation, something in-between). What the discipline actually demands when it talks about plausibilities is rarely obvious.

Consider two quite different things a practitioner might mean by plausible. According to the first reading, a scenario is plausible if nobody in the room can name a reason it could not happen. Call this negative plausibility: plausibility as the absence of identified constraints. According to the second reading, a scenario is plausible only if someone can lay out a causal pathway – a sequence of events, mechanisms, and conditions – running from the present to that “plausible” future. Call this positive plausibility: plausibility as the presence of an identifiable causal route.

Needless to say, the gap between the two is enormous in practice. Negative plausibility is cheap: It depends on ignorance too much, as it refers to us not knowing something – and we know very little. really. The bar is low. The result, if we chose to use this definition of plausibility is that a great many futures count as plausible –all not-yet-refuted. For example, a future in which cold fusion becomes commercially viable by 2040 is negatively plausible in just this sense: nobody has proven it impossible, but nobody can say how it would come about – research is ongoing. Positive plausibility costs more, because it asks the scenario builder to construct, or at least in sketch, a causal path from the present to the described future. Fewer futures can called plausible: no sketch of the path, no plausibility. However, the ones that are plausible under the definition of positive plausibility may well be worth more to a decision-making for the very reason that the causal pathway gives her something to build, monitor, test, and respond to.

The philosophical resources that are to be used here are, we suggest the theory of counterfactuals and the interventionist account of causation. Woodward’s (2003) interventionist account gives a natural way to think about a causal pathway. On Woodward’s account, to identify a cause is to identify a relationship that stays stable under at least some range of interventions. Changing taxation changes consumption, for example. The so-called structural-equation tradition made this concrete (and inspired the interventionist way of thinking about causality). In the causal models of Spirtes, Glymour, and Scheines (1993) and, above all for our purposes, Judea Pearl (2000, 2009), a system is a set of variables tied together by functional relationships, and to evaluate a counterfactual (i.e., contrary-to-fact claim where we say something about something that did not happen) – “what would happen if X took value x?” – one intervenes: Instead of waiting to see what value X happens to take, you reach in, set it to x yourself, disconnect it from whatever normally controls it, and follow the effects down the chain.

Pearl calls this surgery, and a simple example shows why it matters. Suppose I want to know what happens when I turn on a garden sprinkler. I do not worry about why the sprinkler is usually on (maybe a timer, maybe the gardener, whatever). I just switch it on, cut it off from those usual causes (for example, keep the gardener from touching it, no matter how much the person hates to see grass getting too wet), and watch what follows: the grass gets wet, (maybe too wet). This is the difference between doing something and merely seeing it. A wet roof and rain tend to come together, giving us a hunch of causal relationship between the two, but turning on the sprinkler myself tells me what changed as a consequence.

To rely on these very specific (for historical reasons causality has been under extreme stress in philosophy) theories of causality is not to say that every plausible scenario come with a fully specified causal model. That would be unrealistic in futures studies. However, relying on these specific theories tell us that plausibility can and even must carry a causal expectation: that, in principle, a chain of causes joining present to future could be sketched. Otherwise, plausibility is cheap. When a scenario cannot even satisfy this demand for causal sketch that – when nobody can what is being called “plausible” is really “not yet shown to be impossible”. That is a much weaker claim as we saw, too weak and cheap. Honesty in foresight practice means that we strip the name “plausibility” off these claims. One can then go deeper and demand more: that the causal description is not mere sketch, that it is something more – and what this “more” means can be read off from the theories described above about causality. Details, a lot of them are to be found there; but that is not our main topic now.

Adopting positive plausibility as the standard has a cost, then, but it is something we may need to pay. Luckily, there is futures studies literature where work that getting plausibility right, can be seen in the form described above. Cross-impact balance analysis (see Weimer-Jehle, 2006) can be interpreted as operationalising something close to positive plausibility, when it looks for configurations of variables are internally consistent given a specified network of promoting and inhibiting influences, that is, interdependencies that can reasonably, although not uncontroversially, be read as causal ones.

However, positive plausibility is not just a nice term that happens to match some method in futures studies. While general morphological analysis (see Ritchey, 2018) does something similar as cross-impact balance analysis. However, its so-called cross-consistency assessment eliminates combinations that violate known logical, empirical, or normative constraints. In this way, it screens for what can coexist rather than focuses on causal structure as such. Ritchey (2018, 85) is, in fact, explicit that the assessment does not have to refer to causality (even if causal arguments happen to inform a particular consistency judgment). There is no causality here, at least automatically. Thus, there is no plausibility here, at least not automatically. That is: if we follow the definition given for plausibility in this text.

The two methods thus sit, in a sense, on different sides; but they are close at least when it comes to one crucial issue: What they seem to share is a refusal to let mere conceivability settle the question of plausibility or even relevant futures. A concept of plausibility that requires causal pathways would therefore seems not to import a foreign standard but to follow the fields own reasoning. While plausibility is one thing – it is difficult to achieve – there are still methods that are at least close to find plausible futures (even when they bracket causality) and these methods are far from mere speculation, so to say. There is a long continuum between the search for (positive) plausibility and just guessing, that is for sure.

Finally, we wish to pre-empt one worry and possible misreading: that requiring causal pathways biases foresight toward incremental, predictable futures and against the genuinely transformative scenarios we do not know yet how to reach. The worry is important but, in the end, misplaced. First, a causal pathway need not be conventional. It can run through mechanisms that are novel and even unprecedented. What matters is not that the mechanism be familiar but that it be articulable: that the builder of a scenario can say, even roughly, “here is how this could happen.” Second, we are now discussing plausibility. We bite the bullet in the sense that we admit that not all futures are plausible, and only incremental, predictable futures may belong to this group (which is not necessarily to case, see the first reply). To say something is plausible is not to say it is only thing we must take into account when we think about the future. We only wish to give plausibility some posture here – if everything is plausible, then nothing is (and what would then be transformative or unexpected, if plausibility does not serve as contrast?)

Finally, we may notice that Ramírez and Selin (2014) propose an alternative to plausibility itself as a measure of quality of a scenario: productive discomfort and the surfacing of ignorance. They argue that the most useful scenarios sit at an intermediate distance from the familiar – neither so obvious as to be dismissed nor so strange as to be rejected outright – and it is, the argument goes on, there that learning is greatest. This is a high-quality example of a genuine challenge to any framework, ours included, that appears to treat plausibility as a straightforwardly virtuous, at least implicitly. However, we need to distinguish between plausible and positive. One can accept the view about productive discomfort and the surfacing of ignorance as real criteria, but the view does not exclude (positive) plausibility. Plausibility is a requirement on a scenario’s intelligibility – that one can say how it might be causally reached – and says relatively little about the strangeness of the path itself (see also above – paths can be novel). Moreover, a scenario can be causally describable and deeply discomfiting at the same time. One could even argue that disciplining oneself to construct a causal pathway to an unwelcome future is a “good” way to make discomfort productive rather than paralysing. In other words: a scenario being plausible, positively plausible, does not mean the scenario has to be nice or make one satisfied. Plausibility, as a concept, guarantees nothing about positivity. In fact, most plausible futures, like climate changing faster, are least positive.

“Probable”: Interpretation and lessons from the Philosophy of Climate Science

If plausible has been obscure term in futures studies, probable causes even trouble, given the nature of what we have learned about futures thinking. While many like to talk about probable futures, the term has no natural home in this or that tradition. It is used by almost all, even if the notion of probability is far from clear in the phrase probable future.

When we look more closely, we see differences in explicit orientations towards probability. The term probable implies quantification – probabilities, likelihoods, numbers – and futures studies has always had an uneasy relationship with providing exact numbers, given the uncertainties and the basic fact that often the features of the future must be imagined before they can be count. However, some branches of futures studies are more intimately connected to numbers: the Delphi method (see origins in Dalkey & Helmer, 1963; Linstone & Turoff, 1975) often asks panellists at least for probability estimates; cross-impact analysis (Gordon & Hayward, 1968; Helmer, 1981) reasons over conditional probabilities, modeling how the occurrence of one event (re)shapes the likelihood of other events. Other traditions – such as la prospective, CLA, critical futures work in general – are not in the business of providing probability assignments; probability, in futures studies, implies spurious precision, deep uncertainty dressed up as rigour. Both sides have a point about probability as notion in futures studies.

The philosophy of probability can be used to explain both the use and mistrust. The first and most important lesson is that probability is not one concept. An anecdote tells how famous philosopher Russell called it the most important concept in modern science – and then added that nobody has the slightest idea what it means. He was not, if the anecdote is to be trusted, exaggerating. It is a family of concepts. In this family, each concept of probability is grounded in a different conceptual interpretation. When we go further, we notice that each interpretation is appropriate to a different context of application.

The problem is that futures studies rarely pauses to ask which one of the concepts or interpretations of probability is in play in this or that talk about probable futures.

Let’s go to the interpretations, then. The classical interpretation, on the table since Laplace (1814/1999), defines a probability as favourable outcomes divided by all equally possible ones. This works for a dice and shuffled cards, but only because their physical symmetry grounds the fact that the outcomes really are equally likely. Without that type of symmetry, there is no reason to call outcomes equal. Consider the problem through a famous example (see van Fraassen, 1989; after Bertrand, 1889): A cube factory makes cubes up to 2 cm a side, and you ask whether a random dice has a side under 1 cm. Measured by length, the answer looks like half – 50% probability; however measured in terms of volume, it is one part in eight (12.5% probability). Same cubes, same question – but you may receive two answers, shaped only by which yardstick you picked. Futures studies obviously cannot satisfy the symmetry requirements built in this interpretation of probability. A question like “a major disruption in industry and workforce in region X by 2040” has no natural list of outcomes to count, so the favourable-over-total ratio never gets off the ground. And even if you force a list, nothing really makes its entries equally likely – you have simply chosen the items on the list, forced them arbitrarily. This arbitrariness that Bertrand and van Fraassen exposed in their “puzzle”.

The frequentist interpretation treats probability as frequency: how often something does happen as a how often something happens, divided by how many times it could have happened. The finite version counts actual cases: how many members of a reference class have the attribute (Venn, 1876). 7 heads in 10 tosses means a probability of 0.7. The hypothetical version imagines an endless run of trials and takes the frequency it would settle towards (Reichenbach, 1949; von Mises, 1957). The finite version runs straight into what Hájek (1996) calls the problem of the single case, which torments futures studies almost everywhere. Toss a coin a single time and heads comes up either always or never, whatever the coin’s real bias, and therefore the intuitively “real probability” is. events futurists care about (the fall of a particular regime, the arrival of a particular technology, a pandemic from a new pathogen) happen once, if at all – and that is exactly what makes them interesting. This means the frequentist interpretation hardly is meant when futures studies mention probable futures. Futures are tossed at most once: there is only one history. Both versions (the finite and the hypothetical one) also share the reference-class problem discussed above: any event belongs to many classes at once, each with its own frequency. For example, a coming energy transition, the probability of which futures studies may be interested in, might be grouped together with past energy transitions, with infrastructure overhauls in general, or with policy-driven technology shifts (or what have you – the problem becomes greater, the more items you list) and each grouping gives a different number for probability, if taken as frequentist concept.

The subjective or Bayesian interpretation treats probability as a rational agent’s degree of confidence. The tradition was shaped decisively by Ramsey (1926/1990) and de Finetti (1937/1980), who “measured belief” through kind of betting: your degree of belief in an event is the price at which you would buy or sell a bet on it. However, the so-called Dutch Book argument shows the danger in confidence levels that are not coherent. Suppose, in a two-horse race, you call each horse 60% likely to win: the figures each feel reasonable, yet they sum to 120% for an event that must come to exactly 100%. Incoherence of this kind (and it takes subtler forms too, of course, above we have a toy-example) lets a bookie hand you a set of bets you would each accept as fair but that together guarantee a loss (Skyrms, 1984).

The issue is that such incoherencies can be already quietly at work every time a Delphi panellist clicks a percentage button – take, for example 10 claims; who really counts their total percentage? Even if the events (that the Delphi claims present) are then argued to be not mutually exclusive, this is ad hoc and does not really remove the conceptual core of the problem. Moreover, Orthodox Bayesianism, in de Finetti’s style, asks only two things of rational confidence: obey the probability calculus, and update by conditionalisation. This is rather permissive. Two equally rational panellists, given the same evidence, can each assign very different probabilities for the same event. If probable in futures studies means “believed likely by some rational agent,” the notion of probable carries less weight than one might hope. Many rational agents, inflation of probable futures – we guess no one would happy about such inflation.²

The propensity interpretation treats probability as an objective physical disposition or tendency (Popper, 1957, 1959; Peirce, 1957), and it might seem more promising for futures studies: it grounds probability in the items of the world. A society characterized by inequality might be said to have a propensity toward political instability, in the same way a loaded die has a propensity to land on certain faces. But propensity interpretations can be slippery. For example, Hitchcock (2002) notes that calling something a “propensity” does little to specify what the property actually is. Moreover, Humphreys (1985) shows that propensities do not fit Bayes’ theorem, the rule for revising a probability in the light of new evidence (for example, you have a hunch, something happens, and you revise the belief behind the hunch in light of what happened – Bayes’ theorem is simply the exact arithmetic for that revision). The theorem works in both directions, linking the chance of A given B to the chance of B given A, but a propensity only points one way. A viral mutation can have a tendency to cause a pandemic, but it makes no sense to say a pandemic has a tendency to have been caused by that particular mutation – causes push forward to their effects, not backward. The forward probabilities are there; the backward ones the theorem needs are not. This may seem something we can simply ignore as a formal game, but if you buy a interpretation of probability, you must take all of it. And the problem above does matter to futures studies: wherever one reasons backward from a future outcome to its likely cause (how probably a coming pandemic would trace to a lab accident, say), and this is exactly the type of inference propensities cannot make sense of.

What, then, should futures studies do with probability? Maybe the best way to approach the issue is by analogy. The philosophy of climate science offers a good model for handling probability under conditions of deep uncertainty – and it is worth focusing on, because climate scientists face problems also discussed in futures studies. They must assign probabilities to outcomes – global temperature rise, sea-level change, forest fires, heat-waves – that are unique, that arise from complex systems characterized by deep uncertainty, and that are extremely urgent for policies. Climate scientist cannot run the climate system repeatedly to search frequencies. They cannot lean on naive subjective estimates, because the stakes demand intersubjective accountability. And they cannot avoid probability altogether, because decision-makers demand estimates of probability (to count the cost of something against the risks, for example).

In this context a more disciplined form of expert judgment has emerged, one that distinguishes among the sources of uncertainty rather than collapsing them into either false precision or unhelpful vagueness. Its most visible public expression is the IPCC’s calibrated language (see Mastrandrea et al. 2011), which pairs graded probability terms with a separate scale for the strength of evidence and the degree of expert agreement. Rather than asking “how likely is this outcome?” in the abstract, researchers ground their assessments in causal models. In philosophy of climate science Hannart et al. (2016) have built a causal counterfactual framework for the attribution of weather and climate-related events, drawing explicitly on Pearl’s (2000; 2009) interventionist semantics (see above). Stott et al. (2016) argue that pairing physical models with observational constraints allows the factual climate to be compared against a counterfactual world without human “intervention”, which provides estimates of the human contribution to the risk of an event. In this way, the question shifts from “what is the probability of this outcome?” to “how much more, or less, likely does this outcome become given a specific causal intervention?” The move connects probability to the interventionist causal reasoning that, we argued above, should underwrite plausibility. Probability and plausibility are not the same notion, but we do not need to be exact how their difference is formulated. Both can be translated to the language of counterfactual causality. This not only makes the notions easier to grasp but supports our argument that counterfactual reasoning is the “hidden engine” of futures thinking, at least in plausibility and probability.

The lesson for futures studies is not that the field should import climate science’s specific methods as such – we build on an analogy. The lesson is that futures studies should take seriously the distinction between different grounds for a probability assignment. A probability derived from a well-specified causal model, even a rough one, carries more epistemic weight than a probability derived from an someone’s “gut feeling” or “intuition” – even when both are expressed with the same numerical value.

The scenario-planning literature is, in a sense, parallel to this at places at least. Ramírez and Selin (2014) argue that, whatever role probability may play within scenario work, it cannot legitimately serve as something that crowns one scenario the “most likely” to the extent that the scenario is to become a type base case (a practice Ramirez and Selin find in some prominent consulting guidance). This is because the very situations that call for scenarios are situations of Knightian uncertainty (i.e., risk without measurable or quantifiable probabilities; Knight, 1921), where the requisite data and stable distributions simply do not exist. To stamp probability to something under those conditions is not rigour. This is the same caution our distinction between different epistemic grounds (see above) is meant to argue for: a number that no model or adequate real-world data can back up is just a personal estimate dressed up to look like a measurement – and being honest about which it really is makes all the difference. Personal estimates can be extremely useful and “correct” but the difference, on a conceptual level, must be made.

The upshot for the aims of conceptual engineering is the following: When futures studies project or task uses the term probable, it should be clear about which kind of probability is supposed to be at play. At a minimum, practitioners should distinguish between model-based probabilities (derived from specified causal or structural models, bracketing for now their different qualities), expert-judgment probabilities (collected through structured processes like Delphi, with the reasoning documented, not just “clicked”), and subjective credences (individual degrees of confidence, acknowledged as such). These are not equally authoritative, epistemically speaking, and pretending otherwise is where the trouble starts. A futures studies project or task that reports a future as probable while specifying which kind of probability is in question within the claim already represents a real advance in rigour. And when no probability can responsibly be defined – the uncertainty too deep, the models too inadequate, the situation too novel – the honest response is to say so, rather than to disguise ignorance in terminology of probability or likelihood.

“Desirable” and the Hidden Modalities of Preference

The fourth term that deserves attention here and is included in the standard vocabulary is desirable, (often preferable)³ – is usually treated differently than the notions discussed in this text thus far. Possibility, plausibility, and probability look like ontic or epistemic concepts: they describe the world or our knowledge of it and confidence. Desirability is an axiological (value-related) concept: it describes what we find good or just. Bell (1997) dedicated the entire second volume of his Foundations of Futures Studies to normative questions- Bell argued that value judgments are an inescapable part of futures studies. This is true to a great extent. We often ask about desirability. Moreover, to mention some examples, backcasting (Robinson, 1982) takes a desirable future as its starting point and works backward to find the steps to reach it; and there is the “desirable” zone in the futures cone (Voros, 2003) – an area of futures that can be mapped.

However, desirable does not separate from the other modal concepts in its core. To judge a future desirable is, under the surface, to judge it, first of all, possible (assuming one does not rationally desire what one believes to be truly impossible). Second, there are several desirable futures (hopefully) and this means not all of them can actualize, thereby, some are merely possible. Moreover, it seems that to select a desirable future as the target of a backcasting process is to judge it, at minimum, plausible (given our definition of plausible: something that can be reached). In backcasting, the desirable future must be reachable via some pathway from the present. Desirability, in other words, carries hidden modal commitments. Every desirable future implicitly refers possible, sometimes at least plausible, and sometimes achievable, given the right actions.

This means that the quality of a normative futures studies research more than likely depends on the quality of the modal reasoning behind it, at least to the extent the modal reasoning constitutes the researcher at hand. A desirable future that turns out to be physically impossible is not really desirable in any action-guiding sense. Rather, it is a fantasy. A desirable future that lacks any identifiable causal pathway from the present is an ideal not tied to real strategy. When conceptually engineering desirable futures we should therefore drag these hidden commitments into the light: when future studies discusses desirable futures, it should be able to say not only why that future is valued but in what way it is possible, plausible, probable, and so on. The take away lesson is that one cannot simply talk about desirability without committing some other modal concepts. Desirability cannot fly free from any other considerations of what we consider as modal in any sense of that word. There is no one-to-one connection between desirability and some another modal category; sometimes desirable futures may be limited to plausible ones, sometimes even “merely” possible count, and so on. But one has to explicate what connection one means by her use of the notion desirability.

We may finally notice that desirability is connected to agency. In Belnap et al.’s (2001) STIT framework, roughly, each choice available to an agent at a moment narrows the branching futures down to a subset, and an agent “sees to it that” something happens when the outcome is guaranteed by that choice rather than coming about anyway, independently of it. Notice, again, that there is dependency-relations in play here. A desirable future, then, in STIT terms, is one whose realisation depends on some agent’s choices that lead to a better option that would be reached without that choice. If no identifiable agent has the capacity to bring about certain “desirable future”, calling it “desirable” is misleading, wishful thinking. At least in cannot be setting a target. So discussion of ontology of possibilities and modal vocabulary are, not surprisingly, intimately connected. However, we cannot map be the connections here to their full extent.

Counterfactual Reasoning: The Hidden Engine of Scenario Construction

In this writing, we have argued that plausibility should require an identifiable causal pathway and that probability should, where it can, be grounded in causal models – or at least then the probability judgements become most robust. Both requirements share the same type of reasoning that runs through futures studies: counterfactual reasoning. Counterfactual reasoning is disciplined asking of “what-if-things-had-been-different” questions (in its simplest form: what would happen if this variable changed, for example if this or that event occurred, if this policy were adopted, and so on). In what follow, we make more explicit why there are good reasons to consider it as the hidden cognitive engine that drives scenario construction, modal assessment (that is, assessments of possible, plausible, probable, and so on), and most of the analytical work that separates competent foresight from speculation.

The idea is that knowledge of what is possible grows out of our capacity for counterfactual reasoning, i.e., reasoning of the form “if X were the case, then Y would be the case.” We can now put this to work on what scenario builders are actually doing when they construct and evaluate scenarios. A useful guide here is Virmajoki’s (2023) account of causal explanation in history, which discusses, in its final chapter explicitly the similarities between counterfactual based causal explanations if history and futures studies.

The starting point is a simple conception of cause: a cause of an outcome is something that makes a difference to an outcome. A long tradition running back to at least Max Weber (1949), sees causal analysis as the work of separating the factors that made a difference from those that did not. This is also what futures studies does when it the driving forces and critical uncertainties that affect the future – what would change, if this or that was to change? If nothing, there is no causal relevance. If something changes, we have identified something valuable. A good scenario, like a good historical explanation, isolates the variables that are relevant to historical trajectories (including trajectories towards the future; they will become history at some point).

A counterfactual account of (explanation-seeking) reasoning can be based on Woodward’s interventionist idea that an explanation answers what-if-things-had-been-different questions (2003). Explanations are contrastive: they answer “why X rather than Y?” by identifying factors such that, had things been different in some specific way, Y rather than X would have been the case. Scenarios share this superstructure. We may ask “what if oil prices will double rather than remains the same?” – a what-if question – and answer by providing a scenario (or set of scenarios) where we present the consequences in the form that makes clear how the world could have been different from how (the “rather” part) it may be when the prices do not double. In general, a scenario-builder considers a change and traces what would follow.

The interventionist account also gives a clean solution to the problem that is present whenever we discuss the future: which features of the present we hold fixed and what do we let change? If “everything can change” is on the table (meaning that we are allowed to consider that everything changes at once), then chaos follows (of course, we do not argue against the possibility that everything changes, but, we wish to point out that assuming that everything changes suddenly leaves little room to stay on track what the future would be like). In the interventionist (counterfactual) framework, we fix certain factors and ask what follows, if they are thus fixed. To simplify a bit, we stipulate that something is in way or another (or will be) and then follow through. This idea (that certain factors are fixed) is standard in causal explanation literature, and should generate no further trouble for futures studies either. We must specify what changes we allow in any given exercise, and may bracket others for later use.

Consider now the fact, discussed through this text, that how we understand plausible, probable, and so on, differ greatly in their consequences of how seriously we should take the associated judgements concerning plausible or probable (or what have you) futures. This is where a link from history to the future, already present in the earlier work on history-future links (see Staley, 2002; Bradfield et al., 2016 – we do not claim to have discovered the link), becomes most directly relevant to understanding what makes some futures studies better than others. The factors that give the deepest counterfactual explanations – the factors that would have produced similar outcomes across many alternative pasts – are exactly the factors robust enough to matter across many different cases, including possible futures. For example, if alliance systems and geopolitical tensions would have led to large-scale war down numerous alternative historical paths in early 20th century, we have reason to expect that alliance systems and geopolitical tensions to shape the difference between war and peace also in the future. Past contingencies do not repeat in detail, but certain structural causal features seem to repeat themselves in certain forms. It also follows that the sheer number of things that could change a future outcome should not make us unable to say anything about the future: as with the past, not all of those “could”-changes are equally relevant (what if asteroid hit the earth is different than from asking what if alliances were different in early 20th century, to give a toy example), and the work is to separate plausible or probable departures of things from business-as-usual from far-fetched ones. What counts as plausible or probable, of course, depends on how we define these – and the relevance of probable and plausible departures may vary accordingly. Still some departures, some what “what if”:s are more relevant than other; some are plausible, some are merely noncontradictory.

Finally, we may notice how counterfactual reasoning as the engine of scenario construction has several positive consequences for understanding futures studies. First, it explains what makes some scenarios better than others: better scenarios are based on better what-if reasoning, A sharp and robust identification of difference-makers is shared by good counterfactual thinking and futures thinking. Good what-if thinking, in turn, is based on the fact that sometimes we, or some group, have better knowledge of certain domain that the what-if reasoning is about. Experienced practitioners produce better scenarios not because they are more imaginative but because they reason about dependencies with more precision. Or so it seems natural to argue.

Second, it clarifies the relationship between evidence and scenarios. As Virmajoki (2023) notes, historical explanation and evidence are developed hand-in-hand. Explanations cannot be simply extracted from pre-given facts, since deciding what would count as relevant evidence is itself shaped by the explanation under construction. We can know what evidence to look for only once we attempt to sketch an explanation. Scenarios work the same way: The futures we sketch as possible shape which evidence about the present (and past) we treat as evidence for the futures we sketch – and from where we look for the evidence. In general, we should be aware of the relationship between (re)presentations of the future and the evidence they rely on, and how the two are developed and found hand-in-hand.

Third, and most importantly, the idea of counterfactual reasoning as the engine of futures thinking connect the field’s modal vocabulary (possible, probable, plausible, desirable, and so on) in ordinary cognition. We can judge a future possible because we can coherently develop the idea of a condition that would lead to it; we can judge future plausible when we can specify a causal pathway (series of counterfactual dependencies) connecting it to the present; and more or less probable by asking how robust the chain is to our knowledge is – what we consider could change it. A future is desirable if it would be good or just and at least possible. The more disciplined the reasoning in any given case, the better the futures studies. Asking what-if-things-had-been-different questions is at the heart of futures thinking, our argument goes. The issue runs even deeper: To add an icing on the cake, according to Williamson (2007), our knowledge of not just ordinary possibilities but also about metaphysical possibility and necessity is not the product of a special faculty but is simply a by-product of our ordinary capacity to evaluate counterfactual conditionals. Counterfactual reasoning pierces through our minds.

From Vocabulary to Practice

The arguments of this text have put on the table and suggested certain proposals for how the field’s core modal terms should be understood. These are not through-and-through one-size fits for all, imposed, definitions but ideas (building on the practice of conceptual engineering) about refinements and added rigour to what is needed from modal vocabulary of the field. We argued as follows:

Possible should be defined whenever used: at minimum, futures studies should distinguish logical possibility (no contradiction), physical possibility (a future situation that is consistent with laws of nature), and practical possibility (a future situation consistent with laws of nature and other facts we know about how world works); things can be “possible” in many ways, and this matter above all when futures studies is connected decision-making. In the case of plausible, we have taken a bit more explicit stance on how it should be understood: we should require positive plausibility, which means the identification, or at least of a sketch, of a causal pathway from the present to the described future. Negative plausibility, which means the mere absence of identified constraints, is too weak to carry the weight we wish to give to plausible futures in futures studies. Probable can be understood in many ways, and it is difficult to pinpoint the “best” interpretation of the concept. We argued that transparency about the grounds of the assignment: model-based, expert-judgment, or subjective credence (“confidence”). Finally, one must understand that desirable implicit modal commitments that need to be made explicit – one has to tell if a desirable future at hand at any moment is associated with possible, plausible, or probable reading of the modality of that desirable future.

How do these proposals connect to existing futures studies methods? This question is what ties, in conceptual engineering, concepts to practice. There are clear connections between the definitions above and the methods, already discussed (but, of course, we cannot explicate all of them and gaps are possible, to be honest). Horizon scanning and weak-signal (or future sign) detection (Ansoff, 1975; Hiltunen, 2008) can be associated with forms of possibility: they look for and catch signals that indicate that something previously unthought may be possible. The positive plausibility standard does not constrain these methods, since horizon scanning is precisely an attempt to widen the boundary of what we recognise as possible, not to find paths to the possibilities. It is after the signal that the standard bites, if it is to bite: moving from “this might be possible” to “this is plausible” is the moment a causal pathway is sketched.

Scenario planning (see e.g., Wack, 1985a, 1985b; Schwartz, 1991; Schoemaker, 1995; Bradfield et al., 2005; Amer et al., 2013) is perhaps best understood at in terms of plausibility and, in some versions, probability. In our analysis. futures studies thinking is embedded in counterfactual reasoning: the scenario builder considers changes to key variables and traces the consequences. The positive plausibility standard requires that each scenario is delivered with at least an outline of the causal pathway from the present to the future the scenario describes. Skilled practice already tends to do this, but making the requirement concerning causal pathways explicit gives a quality criterion to scenario planning that can be taught, assessed, and improved. Again, we are not saying that something has to be done; we only attempt to set the ground for what could be done.

The Delphi method (long tradition in futures studies, starting from e.g., Dalkey & Helmer, 1963; Linstone & Turoff, 1975;) is perhaps best understood at in terms probability and should be explicit about which interpretation of probability it uses. Notice a trap here: probable is not always the “this is the most automatic to happen; more than plausible (no matter what items like the futures cone presents). The weight of probability assignment depends on how probability is understood and operationalized – that is exactly our argument and where philosophy becomes useful; probability is one topic which is difficult and confusing, and probable cannot be used without care. Be that as it may, the Delphi method is usually built to extract expert-judgment about probabilities. However, an expert who merely clicks “percentage buttons” with no documented reasoning delivers a subjective credence at best (and this subjective credence can often be incoherently structured, strictly speaking; see above). Recording the reasoning, not just the number, lets a Delphi exercise separate well-grounded judgments from guesses and weight them accordingly. Interestingly, Landeta’s (2006) review makes a similar point: Delphi can be taken as valid when applied with methodological rigour, but we must remain cautious if consensus is mistaken for truth and its measurement done without sufficient care. We argue that flagging the grounds of each assessment is one cheap way of keeping the two apart.

We provide, as an example, the observation that cross-impact balance analysis (Weimer-Jehle, 2006) aligns somewhat naturally with positive plausibility, since it looks for configurations of variables that are internally consistent, given certain network of influences; interdependencies that can reasonably be read as causal (causality is tricky even if we accept it as something living inside a method, see e.g., Woodward 2003. In this context, plausibility means the scenario corresponds to a self-consistent configuration. Morphological analysis (Ritchey, 2018) does something similar, but what it calls cross-consistency assessment diagnoses what can coexist rather than causal structure as such. As we saw, Ritchey (2018, 85) is explicit in that the method does need not refer to causality. This means that surviving cross-consistency analysis is not, on its own, positive plausibility in the sense we defended in this text. The conceptual study of this text makes visible what these methods do while steering clear about where consistency stops and causality begins; where plausibility is bracketed, where it is not.

Finally, we may notice how Causal Layered Analysis (CLA) (Inayatullah, 1998) operates at several levels of depth at once, and modalities are intertwined with several issues and work in several ways. No single modal category applies across the levels of CLA uniformly; the demand each level places on a modal claim shifts as one moves through the levels. At the litany level of trends and headlines, the relevant modality comes to close to practical possibility and, where numbers attach, probability: claims here are continuous with business-as-usual and tied to what people think world turns out to be, given what appears obvious. At the systemic level of structures and causes, we move closer to positive plausibility: the analyst identifies causal relations that would have to be made intervention to in order to world to change. As we move onto the level of discourse and worldview, what shifts are the framings and claims of legitimacy that determine which futures possible to think, given one has certain worldview. For example, futures of science are seen differently given different philosophies of science, possibilities open and close as one accepts this or that philosophical worldview on science (see Virmajoki 2022). At the deepest level of myth and metaphor, the modalities reach furthest, probing the logical and metaphysical possibilities that look toward the edge of what we can think, and the standards of modality rightly relax (which does mean this level is easy to map, far from it). Relaxation is not abolition: we can still be precise what type of possibilities we talk about – even plausibilities, given we find mechanism where myth produces visible outcomes or commitment (see discussion in Virmajoki 2022). CLA thus does not escape the modal vocabulary. Rather, it stratifies it, and tracking down the relationship between the levels of CLA and modalities in futures studies would be a worthwhile exercise for that very reason. Unfortunately, that must wait treatment outside the scope of this text.

Conclusion

This writing has argued that the modal vocabulary at the heart of futures studies – the terms practitioners use to classify, weigh, and communicate futures – is in need of engineering, not just interpretation, as the usage is, to put it bluntly, wild. Possible, plausible, probable, and desirable are not mere labels to be applied however one wishes, at least if we wish to maintain some posture in the unity of futures studies and how it communicates to society out there. These modal notions are associated with epistemic claims, and they carry both practical and philosophical weight. That weight can only be assessed, if modal vocabulary is communicated clearly. Possibility should specify what kind of possibility. Plausibility should say whether it is associated with causal pathways – if not, we argue, it is term used too loosely. Probability should perhaps not deeply disclose its exact interpretation but still explain what grounds when one is saying something is probable. Desirability claims should make explicit their hidden commitments to other forms of modality discussed in this text.

We have also argued that counterfactual reasoning is the hidden cognitive engine of scenario construction and, in general, of the futures studies when it moves in the space of possibilities – what could have happened, had this or that changed is structurally similar to what may happen, given this or that change. The claim about the hidden engine ties well together, when looked closely, with the work in futures studies – asking “what if?”, tracing consequences, evaluating pathways – which is why we make the argument in the first place. Understanding conceptual core is necessary, and our claim about counterfactual reasoning is a claim about this conceptual core, when it comes to modality. The link is not decorative, then. Counterfactual reasoning is to foresight what calculation is to accounting; it is not the whole of the practice, but it is operation without which the practice could not get off the ground. Counterfactual armchair pondering is no substitute for real methods.

But sharpening the vocabulary raises a further question we have raised but not really answered. If plausibility requires causal pathways and probability requires transparency about grounds, what counts, in detail, as relevant evidence for such claims? How do we know that a causal pathway is genuine and not just a plausible-sounding story – causality is tricky, as noted (and often loosely used as a concept in futures studies, unfortunately – where are, say, interventionist causal models or equations beyond forecasting, where talk about causality is still there)? How do we assess whether probability assignments are well calibrated and coherent? What are the limits of our ability to know what lies ahead at all – what really determines what we can think about possibility even in its widest sense? These are epistemological questions – questions about the nature and the limits of the knowledge that futures studies produces and can produce. This raises another challenge: what, exactly, can be known about the future, and how? No matter how much effort we spend to conceptually engineering modal vocabulary, the work is empty without associated treatment of the knowledge that is transmitted with and through modal vocabulary. Again, this is out of the scope of this writing.

Bibliography

Amara, R. (1981). The futures field: Searching for definitions and boundaries. The Futurist, 15(1), 25–29.

Amer, M., Daim, T. U., & Jetter, A. (2013). A review of scenario planning. Futures, 46, 23–40.

Ansoff, H. I. (1975). Managing strategic surprise by response to weak signals. California Management Review, 18(2), 21–33.

Armstrong, D. M. (1997). A world of states of affairs. Cambridge University Press.

Bell, W. (1997). Foundations of futures studies: Vol. 2. Values, objectivity, and the good society. Transaction Publishers.

Belnap, N., Perloff, M., & Xu, M. (2001). Facing the future: Agents and choices in our indeterminist world. Oxford University Press.

Bertrand, J. (1889). Calcul des probabilités. Gauthier-Villars.

Blackburn, S. (1999). Think. Oxford University Press.

Bradfield, R., Wright, G., Burt, G., Cairns, G., & van der Heijden, K. (2005). The origins and evolution of scenario techniques in long range planning. Futures, 37(8), 795–812.

Bradfield, R., Derbyshire, J., & Wright, G. (2016). The critical role of history in scenario thinking: Augmenting causal analysis within the intuitive logics scenario development methodology. Futures, 77, 56–66.

Brun, G. (2016). Explication as a method of conceptual re-engineering. Erkenntnis, 81(6), 1211–1241.

Burgess, A., & Plunkett, D. (2020). On the relation between conceptual engineering and conceptual ethics. Ratio, 33(4), 281–294.

Cappelen, H. (2018). Fixing language. Oxford University Press.

Cappelen, H., & Plunkett, D. (2020). Introduction: A guided tour of conceptual engineering and conceptual ethics. In A. Burgess, H. Cappelen, & D. Plunkett (Eds.), Conceptual engineering and conceptual ethics (pp. 1–26). Oxford University Press.

Carnap, R. (1950). Logical foundations of probability. University of Chicago Press.

Carus, A. W. (2007). Carnap and twentieth-century thought. Cambridge University Press.

Chalmers, D. J. (2002). Does conceivability entail possibility? In T. S. Gendler & J. Hawthorne (Eds.), Conceivability and possibility (pp. 145–200). Oxford University Press.

Chalmers, D. J. (2025). What is conceptual engineering and what should it be? Inquiry, 68(9), 2902–2919.

Creath, R. (1991). Every dogma has its day. Erkenntnis, 35, 347–389.

Dalkey, N., & Helmer, O. (1963). An experimental application of the Delphi method to the use of experts. Management Science, 9(3), 458–467.

Dator, J. (2009). Alternative futures at the Manoa School. Journal of Futures Studies, 14(2), 1–18.

de Finetti, B. (1980). Foresight: Its logical laws, its subjective sources (H. E. Kyburg Jr., Trans.). In H. E. Kyburg Jr. & H. E. Smokler (Eds.), Studies in subjective probability (pp. 55–118). Robert E. Krieger. (Original work published 1937)

Dutilh Novaes, C. (2020). Carnapian explication and ameliorative analysis. Synthese, 197(3), 1011–1034.

Floridi, L. (2011). The philosophy of information. Oxford University Press.

Gordon, T. J., & Hayward, H. (1968). Initial experiments with the cross-impact matrix method of forecasting. Futures, 1(2), 100–116.

Hájek, A. (1996). “Mises redux”—Redux: Fifteen arguments against finite frequentism. Erkenntnis, 45, 209–227.

Hale, B. (2013). Necessary beings. Oxford University Press.

Hancock, T., & Bezold, C. (1994). Possible futures, preferable futures. Healthcare Forum Journal, 37(2), 23–29.

Hannart, A., Pearl, J., Otto, F. E. L., Naveau, P., & Ghil, M. (2016). Causal counterfactual theory for the attribution of weather and climate-related events. Bulletin of the American Meteorological Society, 97(1), 99–110.

Haslanger, S. (2005). What are we talking about? The semantics and politics of social kinds. Hypatia, 20(4), 10–26.

Haslanger, S. (2012). Resisting reality: Social construction and social critique. Oxford University Press.

Helmer, O. (1981). Reassessment of cross-impact analysis. Futures, 13(5), 389–400.

Hiltunen, E. (2008). The future sign and its three dimensions. Futures, 40(3), 247–260.

Hitchcock, C. (2002). Probability and chance: Philosophical aspects. In International encyclopedia of the social & behavioral sciences (Vol. 18, pp. 12089–12095). Elsevier.

Humphreys, P. (1985). Why propensities cannot be probabilities. The Philosophical Review, 94(4), 557–570.

Inayatullah, S. (1998). Causal layered analysis: Poststructuralism as method. Futures, 30(8), 815–829.

Isaac, M. G., Koch, S., & Nefdt, R. M. (2022). Conceptual engineering: A road map to practice. Philosophy Compass, 17(10), Article e12879.

Knight, F. H. (1921). Risk, uncertainty, and profit. Houghton Mifflin.

Koch, S. (2021). There is no dilemma for conceptual engineering: Reply to Max Deutsch. Philosophical Studies, 178(7), 2279–2291.

Kripke, S. A. (1980). Naming and necessity. Harvard University Press.

Landeta, J. (2006). Current validity of the Delphi method in social sciences. Technological Forecasting and Social Change, 73(5), 467–482.

Laplace, P. S. (1999). A philosophical essay on probabilities (A. I. Dale, Trans.). Springer. (Original work published 1814)

Lewis, D. (1980). A subjectivist’s guide to objective chance. In R. C. Jeffrey (Ed.), Studies in inductive logic and probability (Vol. 2, pp. 263–293). University of California Press.

Lewis, D. (1986). On the plurality of worlds. Blackwell.

Linstone, H. A., & Turoff, M. (Eds.). (1975). The Delphi method: Techniques and applications. Addison-Wesley.

Mastrandrea, M. D., Mach, K. J., Plattner, G.-K., Edenhofer, O., Stocker, T. F., Field, C. B., Ebi, K. L., & Matschoss, P. R. (2011). The IPCC AR5 guidance note on consistent treatment of uncertainties: A common approach across the working groups. Climatic Change, 108(4), 675–691.

Nado, J. (2021). Conceptual engineering, truth, and efficacy. Synthese, 198(Suppl. 7), 1507–1527.

Nado, J. (2023). Classification procedures as the targets of conceptual engineering. Philosophy and Phenomenological Research, 106(1), 136–156.

Pearl, J. (2000). Causality: Models, reasoning, and inference. Cambridge University Press.

Pearl, J. (2009). Causality: Models, reasoning, and inference (2nd ed.). Cambridge University Press.

Peirce, C. S. (1957). Notes on the doctrine of chances. In V. Tomas (Ed.), Essays in the philosophy of science (pp. 74–84). Bobbs-Merrill.

Plantinga, A. (1974). The nature of necessity. Oxford University Press.

Popper, K. R. (1957). The propensity interpretation of the calculus of probability and the quantum theory. In S. Körner (Ed.), Observation and interpretation (Colston Papers, Vol. 9, pp. 65–70). Butterworths.

Popper, K. R. (1959). The propensity interpretation of probability. The British Journal for the Philosophy of Science, 10(37), 25–42.

Priest, G. (2021). Metaphysical necessity: A skeptical perspective. Synthese, 198(Suppl. 8), 1873–1885.

Queloz, M., & Bieber, F. (2022). Conceptual engineering and the politics of implementation. Pacific Philosophical Quarterly, 103(3), 670–691.

Ramírez, R., & Selin, C. (2014). Plausibility and probability in scenario planning. Foresight, 16(1), 54–74.

Ramsey, F. P. (1990). Truth and probability. In D. H. Mellor (Ed.), Philosophical papers (pp. 52–94). Cambridge University Press. (Original work published 1926)

Reichenbach, H. (1949). The theory of probability. University of California Press.

Ritchey, T. (2018). General morphological analysis as a basic scientific modelling method. Technological Forecasting and Social Change, 126, 81–91.

Robinson, J. B. (1982). Energy backcasting: A proposed method of policy analysis. Energy Policy, 10(4), 337–344.

Schoemaker, P. J. H. (1995). Scenario planning: A tool for strategic thinking. Sloan Management Review, 36(2), 25–40.

Schwartz, P. (1991). The art of the long view: Planning for the future in an uncertain world. Doubleday.

Shields, M. (2021). Conceptual domination. Synthese, 199(5–6), 15043–15067.

Shoemaker, S. (1998). Causal and metaphysical necessity. Pacific Philosophical Quarterly, 79(1), 59–77.

Skyrms, B. (1984). Pragmatics and empiricism. Yale University Press.

Spirtes, P., Glymour, C., & Scheines, R. (1993). Causation, prediction, and search. Springer-Verlag.

Staley, D. J. (2002). A history of the future. History and Theory, 41(4), 72–89.

Stott, P. A., Christidis, N., Otto, F. E. L., Sun, Y., Vanderlinden, J.-P., van Oldenborgh, G. J., Vautard, R., von Storch, H., Walton, P., Yiou, P., & Zwiers, F. W. (2016). Attribution of extreme weather and climate-related events. WIREs Climate Change, 7(1), 23–41.

Strohminger, M. (2015). Perceptual knowledge of nonactual possibilities. Philosophical Perspectives, 29(1), 363–375.

Thomasson, A. L. (2025). Conceptual engineering: When do we need it? How can we do it? Inquiry, 68(9), 2993–3018.

Vaidya, A. J., & Wallner, M. (2021). The epistemology of modality and the problem of modal epistemic friction. Synthese, 198(Suppl. 8), 1909–1935.

van Fraassen, B. C. (1983). Calibration: A frequency justification for personal probability. In R. S. Cohen & L. Laudan (Eds.), Physics, philosophy and psychoanalysis: Essays in honor of Adolf Grünbaum (pp. 295–319). D. Reidel.

van Fraassen, B. C. (1989). Laws and symmetry. Clarendon Press.

van Inwagen, P. (1998). Modal epistemology. Philosophical Studies, 92(1–2), 67–84.

Venn, J. (1876). The logic of chance (2nd ed.). Macmillan.

Vetter, B. (2023). An agency-based epistemology of modality. In D. Prelević & A. Vaidya (Eds.), The epistemology of modality and philosophical methodology (pp. 44–69). Routledge.

Virmajoki, V. (2022). Understanding futures of science: Connecting causal layered analysis and philosophy of science. Journal of Futures Studies, 27(2).

Virmajoki, V. (2023). Causal explanation in historiography. Palgrave Macmillan.

von Mises, R. (1957). Probability, statistics and truth (rev. English ed.). Macmillan.

Voros, J. (2003). A generic foresight process framework. Foresight, 5(3), 10–21.

Wack, P. (1985a). Scenarios: Uncharted waters ahead. Harvard Business Review, 63(5), 72–89.

Wack, P. (1985b). Scenarios: Shooting the rapids. Harvard Business Review, 63(6), 139–150.

Weber, M. (1949). The methodology of the social sciences (E. A. Shils & H. A. Finch, Eds. & Trans.). Free Press.

Weimer-Jehle, W. (2006). Cross-impact balances: A system-theoretical approach to cross-impact analysis. Technological Forecasting and Social Change, 73(4), 334–361.

Williamson, T. (2007). The philosophy of philosophy. Blackwell.

Woodward, J. (2003). Making things happen: A theory of causal explanation. Oxford University Press.

Yablo, S. (1993). Is conceivability a guide to possibility? Philosophy and Phenomenological Research, 53(1), 1–42.

1. We use the term desirable whenever possible, i.e., when we are not citing author whose use of preferable is not at the very core of their writing.↩︎

2. Bayesians have, of course, added further constraints such as matching credences (confidence) to observed frequencies (van Fraassen, 1983), or the Principal Principle, which says that when you know the real, objective odds of something, your personal confidence should simply match them (Lewis, 1980). However, these constraints tame subjective probability by tying it to something outside the individual mind – to frequencies, or to chances handed down by our best models – and share similar issues as the interpretations based directly to them.↩︎

3. As noted, we use the term desirable whenever possible, i.e., when we are not citing author whose use of preferable is not at the very core of their writing.↩︎

Translation from Finnish learning material to be published on open online pages

Key Points

• Knowledge does not replace values, nor do values replace knowledge. Without knowledge, futures thinking turns into hope or fear; without values, into groping without direction or into submission to the aims of others.
• There can genuinely be many values, but this does not mean that everything is valuable or right. Pluralism without relativism is a difficult but necessary attitude: it makes reflecting on the values of futures more difficult but at the same time more precise and more open.
• Different theories of morality and justice give different answers as to which future is desirable. Utilitarianism, for example, looks at the goodness of consequences, Rawls raises the position of the worst-off, and capability thinking directs its gaze to real opportunities for action and to power relations connected with equality. It is perhaps best to use the theories together and in layers, not to choose a single determiner of desirability.
• Whose values get to define the futures put up for assessment is both an epistemic and an ethical question. When certain perspectives and positions are excluded, the problem is not only ethical but also concerns knowledge: knowledge of futures that are visible only from other positions.
• Values themselves also change – and this has often been a blind spot of futures research. The field often anticipates, justifiably, technological and societal change but frequently assumes the unchangingness of values, even though future generations may look at our current ethical commitments with the same criticism with which we look at past ones.

Some basics of ethics and values

Before values can be discussed in more detail in futures research, it is worth pausing over a few basic concepts. Futures are always evaluated with the help of certain concepts, and if these concepts are not distinguished from one another, they easily begin to steer research and futures thinking unnoticed. Desirability is an ethical concept that is used often, but a precise, or even approximate, definition is rarely brought out in the context of use. Situations then arise in which people seem to be speaking of the same thing, even though one is speaking, for example, of moral obligation, another of the good life, a third of the functioning of society, and a fourth of their own or their organization's goals – all in the name of desirability.

Within the basic conceptual toolkit, morality and ethics must be distinguished. By morality is usually meant those norms, rules, and principles that are thought to guide our action in relation to other persons and other matters. It concerns, for example, what may be done to others and what may not; what we are responsible for; when some act is wrong even if it is useful, efficient, or expedient with respect to some goal. Morality is a weighty matter in life. It is not merely a question of what we happen to find pleasant or meaningful, but of what is considered to obligate, or at least to give a strong reason, to act in a certain way, however this way of acting aligns with our momentary desires.

Ethics, in turn, can be understood as a broader conception of the good life, of right ends, and of what kind of action, community, or society is just. Ethics is not limited to what is forbidden or permitted. It also concerns what kind of life is worth pursuing, what kinds of relations between people are held to be good, and what kinds of ends are such that pursuing them brings good – or at least does not destroy something important. Morality can indeed be seen as resting on a more general ethical level.

Values, in turn, are those things we hold important, worth pursuing, or good in themselves. As general examples we can mention how we value freedom, justice, happiness, knowledge, friendship, security, health, nature, human dignity, and so on.

This division is not sharp-edged; rather, morality, ethics, and values form a bundle in which the different parts shape one another. If my value is happiness, I can take an ethical stance according to which happiness is the factor in whose light morality is assessed. We need not unpack the whole bundle; the essential thing is to notice that judgments of desirability concerning the future often lie between all these levels or are a combination of them. When we reflect, for example, on the future of education, we do not deal only with learning outcomes, funding, or technological solutions. The reflection takes a stand, at least without saying so aloud, on what is regarded as valuable learning (what would be good to learn in order to become part of a functioning society), what obligations society has especially toward children and the young, what opportunities should be secured for people, and what kind of life education is thought to serve. Correspondingly, a sustainable economy is not a question only of the use of resources or the efficiency of production. It contains a conception of moderation, responsibility, justice, and of whose well-being is included in the assessment. All assessments can go ethically astray, but they are always present; we can never hide behind the idea that reality is one thing and ethics another, to be spoken of separately. This is not possible.

In philosophy, morality has traditionally been structured with the help of a few classical approaches. Let it be said at once that from the standpoint of futures research they need not be treated as competing total systems, of which one is chosen and the others rejected. It is more useful to see them as different ways of directing attention to the grounds on which some act, institution, policy, or future more generally can be assessed as good or bad. In this sense they are instruments for making better choices, not battle cries.

Consequentialist ethics assesses acts and solutions on the basis of their consequences. Its best-known form is utilitarianism, in which what produces the most good or benefit overall is held to be morally right. Classically this way of thinking was formulated by Jeremy Bentham and John Stuart Mill, although there are important differences between them, for example in how the quality of pleasure or well-being is understood (see Bentham 1789; Mill 1863) – the central question of utilitarianism is indeed when a consequence is good. Whatever the final answer, in futures work (thinking concerning the future and the series of acts that follows it) this way of thinking is natural, because we often want to know what consequences our acts will have for the future. A problem of sorts is that the use of this perspective is left unsaid. When scenarios are compared on the basis of overall impacts, costs, welfare effects, risks, or benefits, we move close to the mental landscape of consequentialist ethics. Yet thoughts of, for example, aggregate benefit can conceal how, and to whom, the benefits and harms are distributed. A future can look good “at the level of averages” even if it were intolerable for some – and a castle in the air for others.

Duty ethics, especially as commonly associated with Kant's thought, is a different perspective. Not everything can be justified or “valued as good” by good consequences. According to Kant, a human being must not be treated merely as a means but always at the same time as an end in themselves (see Kant 1785). In futures research this observation is important, because many decisions concerning futures can easily be justified by advantages that do not concern the individual but bury the individual beneath them: economic growth, efficiency, security, technological development, or improved crisis resilience. From the perspective of duty ethics, however, one must ask whether some individual is then left in a position we cannot morally accept. Some future can be “efficient” and still wrong if it requires using some people as mere means for the goals of others – efficient, that is, from the standpoint of consequentialist ethics.

Virtue ethics, in turn, shifts attention from acts, their consequences and their limits, to agents and their character. In the (Aristotelian) tradition stemming from antiquity, the question is what kind of person acts morally well, not only what rule or principle should be followed in a given situation (see Aristotle, Nicomachean Ethics). In futures research this may feel distant, but the perspective is in fact intertwined with the field. Futures are built not only by technologies, institutions, and policy measures, but also by the dispositions of agents and by their character and capacity: courage or cowardice, moderation or greed, deliberation or short-sightedness, solidarity or indifference. There is thus reason to examine what kinds of agents a given future requires and what kinds it in turn produces. What are we like – are we virtuous, future-producing agents?

Contract ethics and so-called contractualist ways of thinking bring in questions of acceptability more directly. What is morally significant is then what kinds of principles people could justifiably accept, or what no one could reasonably reject (see Scanlon 1998). Rawls's theory of justice belongs to this broad family (see in more detail below), although it is at the same time its own particular (and of course, as always, contested) theory (see Rawls 1971). For futures research this contractualist perspective is important because futures also concern those who are not present to assess and create them now: children, future generations, other species, or those societal groups that were not invited into the futures work. In the assessment one must then take into account what kinds of principles guiding the future could be acceptable from the standpoint of those who will have to live with these principles. Who is the one who accepts the contract that prevails in the future as well?

From these approaches one need not, at least at this stage, choose the one correct one. It is more important to notice that they direct our thinking to different things. Consequentialist ethics makes impacts and wholes visible, duty ethics limits and respect, virtue ethics the quality of agents, and contract ethics acceptability. In discussion concerning the future these ways of thinking can blend into one another (sometimes meaningfully, to be sure) in such a way that our thinking becomes blurred. We may debate as if it were only a matter of facts, even though the disagreement concerns the grounds on which the goodness of the future should be assessed – facts and values are connected to each other, but a disagreement about facts that rests on a value conflict is unresolvable. On the other hand, it can sometimes also be imagined that a dispute is purely about values, even though the differences of view are connected to how each party sees the facts of reality.

For futures research it is precisely this connection between values and reality that is essential. Values do not replace knowledge, but neither does knowledge replace values. A conception is needed of what can happen, on what conditions, and with what uncertainties, so that what is to come can be reflected on from an ethical perspective. On the other hand, a conception is often needed of why some possibility is held to be worth pursuing, guarding against, or preventing, so that we can understand why we move, or make strategy, toward this possibility. Without reality, futures thinking turns into hope or fear; without values, it turns into verbal assessment without an understanding of the human being as agent and experiencer.

What values are – and why there are many

In everyday language we speak of values rather broadly. We say we value freedom, security, friendship, cultivation and learning, nature, justice, family, knowledge, creativity, or, say, peace. These are not, however, the same kinds of things, and our values indeed attach to different kinds of matters. Some relate to societal institutions, some to relations between people, some to the individual's good life, and some to things we hold valuable regardless of immediate benefit. Yet in a certain way these things must be capable of being distinguished from one another.

In futures research this multiplicity comes up quickly, because the field does not study only certain futures but the futures of different things and topics. When we describe, for example, possible futures, inside these futures there are valuable things of different types. From this follow different combinations of, and conflicts between, values and norms. For example, the future of education concerns competence, cultivation, equality, labor markets, individual freedom, and the continuity of society. Cultivation and the fairness of labor markets are different as values. The future of technology, again, concerns efficiency, control, security, freedom, privacy, and power. Efficiency and security are, again, different values.

In the discussion concerning values, one central question is whether the different values are ultimately reducible to one basic value or not. Monism means the view according to which behind everything valuable there is, in the last instance, one value – for example well-being, pleasure, happiness, or the satisfaction of preferences. Other things are then valuable insofar as they promote this fundamental good or are in some way derivable from it. Then, for example, freedom, knowledge, or friendship are valuable because (or insofar as) they increase well-being, or insofar as well-being requires them.

Value pluralism, in turn, is the view according to which there genuinely are many values. They are not merely different derivations of one value or good, or a precondition of its prevailing. For example, freedom and efficiency can on this view both be real values, even if in certain situations they come into tension – do we want to make a sector more efficient even if this leads to the control of people's time? According to a more classical example, knowledge and happiness can both be important, even though knowing the truth does not always make a person happier.

The strength of value pluralism is that it corresponds rather well to our experience. Moral and political choices often seem to weigh incommensurable things – that is, things whose values cannot be compared on the same scale, such as efficiency and freedom (on incommensurability, see Chang 1997). Often a choice can be justified and still leave behind a meaningful regret that something valuable genuinely had to be given up. This is intelligible only if the rejected alternative was genuinely valuable, not merely a lesser form of the same value. It can be thought that values stand in genuine conflict and that this is not a matter of some misunderstanding of the nature of the other value (see Berlin 1958).

In futures research this has great significance. If all values are reduced to one basic value, assessing the goodness and rightness of futures is seemingly easy. One can, for example, assess which scenario produces the most benefit, growth, or satisfaction, or the least harm – one of these is taken as the basic value and the assessment is carried out. But if there are many values, the assessment of futures requires distinctions. One must assess, for example, in what ways a given future promotes well-being, in what ways it strengthens freedom, and in what ways it treats vulnerable groups. Then one must weigh whether the combination of values that a certain future contains is more desirable than some other.

It must be underlined, however, that what is required is weighing, not blindness. Relativism does not follow from pluralism. It is important to distinguish the multiplicity of values from the idea that all value systems would be equally good. We can perfectly well think that freedom, equality, well-being, and human dignity are separate values, and still hold slavery, arbitrary cruelty, or systematic subjugation to be wrong. Multiplicity does not remove the possibility of criticism. It makes criticism more difficult but also more precise – we do not use an arbitrary mechanism in which we reduce a bundle of values to some basic value that a given future would bring with it.

This distinction is decisive for futures thinking. If pluralism is confused with relativism, futures research loses its capacity to say why some future is better or worse than another. If, on the other hand, pluralism is warded off and all values are, as it were, forced onto one measure, futures research becomes impoverished and begins to see futures in the light of only one value. Both errors are possible but avoidable.

Yardsticks of a desirable future

If futures research assesses the ethicality and desirability of futures, it must in some way take a stand on what makes a given future desirable. This question cannot be resolved at once or finally, but neither can it be avoided. To understand alternative answers, we can look at philosophy's theories of justice. There is no reason to treat them as academic hair-splitting, for the different answers lead to different conclusions about what kinds of futures are held to be worth pursuing. In this section we deal with a few examples.

One of the most common and often unstated starting points is the utilitarian one, which we encountered above. When, for example, an administration assesses which future would enable the most efficient public administration, or when a company chooses a future on the basis of the greatest expected return, they move at least to some degree within utilitarian thinking. The question concerns what benefit some decision might have on a certain measure, such as efficiency (which can of course be reduced to a basic value such as the increase of material well-being). Such thinking makes it easier to assess the future in the midst of uncertainty and conflicting views.

The problem, however, is fairly obvious. In the spirit of utilitarianism we could accept a future in which overall well-being is great but is distributed glaringly unequally. If the gains of the winners are large enough, the losses of the losers can remain to one side in the overall calculation – putting it somewhat pointedly. In addition, utilitarianism can accept people's preferences uncritically. A community accustomed to living in poor conditions can express preferences that have adapted precisely to these conditions. Then a future that would look acceptable on the basis of preferences (because it would increase the things that satisfy these preferences) can in fact be the continuation of injustice in another form (see Sen 1992). In futures research this is especially important, because the preferences of the present moment do not always tell us what would in the future be genuinely good or right for people. We return to the theme in more depth below.

Rawlsian justice thinking offers one influential but contested alternative. John Rawls's well-known thought experiment is the so-called veil of ignorance: the principles of society are chosen as if we did not know what position we ourselves will end up occupying in that society (see Rawls 1971). According to Rawls, in such a situation it would be rational to choose principles that protect the worst-off as well. Everyone must have the same rights, and differences between people are justified only if they rest on a system that improves the position of the worst-off. From this perspective one does not ask only which future produces the most well-being overall, but what kind of future treats its worst-off best, or at least improves their position. The desirability of a future depends on the position of the worst-off.

We could indeed draw a certain analogy to futures work, for we often think of generations in whose position we ourselves are not – although, of course, the mechanisms of intergenerationality make neutrality impossible. The difficulty, however, is that in applying Rawls's ideas one should be able to specify who the worst-off of each future are. How disadvantage is determined is not only a conceptual question; it also involves an empirical part. In a technologically advanced future, disadvantage can relate to competence, to access to systems, or to the capacity to control one's own data – however wealthy a person may otherwise be. As the climate crisis intensifies, it can relate to place of residence, health, or the capacity to relocate. In a security-policy future it can relate to vulnerability, relations of dependence, or the capacity to influence politics. How are the worst-off carved out of the different futures – do we then make a new set of possible futures that focus only on them? From the standpoint of practical research this would be challenging.

The capability approach developed by Amartya Sen and Martha Nussbaum is also worth keeping an eye on. The central idea of the approach is that the same resources do not guarantee the same freedom for everyone. For example, a physically injured person may need more resources to achieve the same actual capacity for action as one who has avoided injury. On the other hand, a person living in a poor environment does not necessarily benefit from formally identical rights in the way a person does who has the preconditions, provided by institutions, education, and infrastructure, for carrying out acts in accordance with their rights. What is essential is thus not only what people have, but what they can actually do and be – what they are capable of (see Sen 2009).

Nussbaum specified this approach by listing central human capabilities, such as life, health, bodily integrity, imagination, emotions, practical reason, affiliation, relation to other species, play, and control over one's environment (see Nussbaum 2000). For futures work this gives a multidimensional way of examining the elements of a desirable future. A future is not desirable, for example, only because we achieve better resources or because some system works more efficiently. One must look at what people's opportunities to act, their capabilities to act, are in that future. Can people live, learn, participate, move, form relationships, influence their environment, and regard themselves as agents? Of course, a problem quickly arises here: if the list of features of a desirable future becomes long, distinguishing between such futures becomes truly difficult – what to take into account, and how to balance if conflicts arise?

According to sufficiency thinking (i.e. sufficientarianism), in turn, the most important thing is not necessarily perfect equality or the maximization of aggregate benefit, but ensuring that everyone crosses a certain threshold of sufficiency. Everyone must have enough. The basic idea is thus that “moral urgency” depends on whether all people get above a certain threshold with respect to a certain matter. For example, the insufficiency of medicines for people on low incomes is not acceptable, because people do not receive sufficient resources (medicines) as a precondition of their well-being. Above the threshold, the moral urgency of further improvements can diminish, even if they do not become indifferent (see Frankfurt 1987). For futures work this gives a usable sieve. If there are two futures, in one of which a certain morally relevant threshold is crossed and in the other not, the first is the desirable future.

Sufficiency thinking has been criticized for not saying enough about injustices above the threshold (see Casal 2007). It is perhaps possible to imagine a society in which everyone has enough (let the definition be taken as given), but in which power, wealth, esteem, or opportunities to exert influence are nevertheless distributed in deeply problematic ways. It is perhaps blunt to ask whether a future in which one person remains below the threshold but the others live nearly identical lives really is better than a future in which even that one person crosses the threshold, but a certain elite has opportunities of which most can only dream. It is perhaps best to regard sufficiency thinking as one condition, not the whole truth, for when a future is acceptable – not, perhaps, desirable. We can accept only futures in which resources suffice, but some of these can be less desirable than others, for example because of a glaring inequality of wealth.

In practical futures work, the theories presented above (and of course others as well – we dealt with only a selection by way of example) can be used together. For instance, one could construct the following filter for desirability: First one examines whether everyone crosses the threshold of sufficiency. When this has been checked, we direct our gaze to what kinds of capabilities the futures thus first screened would enable. We select the capability-increasing futures for our list. After this, we assess what kinds of power relations and forms of subjugation – more generally, inequality – these different futures generate or dismantle, and on the basis of what mechanism. Such a layered approach does not necessarily produce a single ranking. Its value lies in making the normative questions visible and workable instead of resolving them with people's haphazard intuitions.

Whose values decide

The theoretical approaches discussed above do not yet resolve one central question. When assessing on what grounds a given future is desirable, one must examine whose values get to define, in the first place, what kinds of futures are set up for assessment at all. This question concerns the justice of participation (or inclusion). Inclusion is both an epistemic and an ethical question in futures research.

The epistemic side concerns what follows when certain voices are excluded from futures work. It can then be argued that futures research loses access to perspectives that could improve its quality. Miranda Fricker's (2007) concept of epistemic injustice helps in understanding this. A person can be given less credibility than they would deserve because of a prejudice attached to their identity. On the other hand, they themselves may lack the conceptual or institutional means to express their own experience in a commonly intelligible or accessible way. Knowledge remains in the margins of society.

For example, an expert panel consisting exclusively of people of similar education, generation, and societal position can produce internally coherent foresight. This does not yet necessarily make the foresight sufficient. The panel can “systematically” pass over futures that would be visible only from other positions. One sees technological efficiency, another dependence. One sees the automation of a service, another the loss of access. One sees an increase in security, another a thickening of surveillance. We do not, however, want to end up in a situation in which genuine expertise and researched knowledge are regarded as being as significant (or insignificant) as anyone's sketches of reality. Rather, the point is that it is possible that the research itself is already in a certain sense unjust and must therefore also be examined through the lens of justice. Nor is it a matter of values steering research (see the article on conceptions of knowledge) but of how knowledge itself can constitute a question of justice.

Besides epistemic justice there are of course fundamental questions concerning the investigation of values, such as the question of whose values appear thinkable at all. For example, so-called critical-interpretive methods, such as causal layered analysis (CLA), start from the premise that beneath the surface of images of the future there are deeper layers: worldviews, myths, metaphors, and basic assumptions that delimit what is held possible or meaningful to think about the future (see Inayatullah 1998). When we assess the desirability of a future, these layers and the values built into them must be gone through. We cannot simply be fully aware of what values we ultimately carry with us and what they rest on.

Consider, for example, a future built around a value foreign to us, such as (to take a “toy example”) the requirement that dwelling floor area in square meters must be a prime number. From the present, this future can appear indifferent, strange, or downright incomprehensible. This is not necessarily because that future is impossible, undesirable, or indifferent. It may be that our understanding lacks the means to express and recognize the value. Perhaps the prime-number system would enable better mobility, because furniture in that future is even-numbered in area? A silly example, but we cannot in any case demonstrate the idea otherwise (if the example were genuine, it would no longer demonstrate a strange future but something we already comprehend). One task of futures research is precisely to make these limitations visible. Not because the foreign would somehow automatically be better, but so that unnoticed assumptions do not steer the discussion without it being possible to discuss them.

What is essential, then, is not only who is invited to the table, but also whether the conceptual frames concerning futures are open to discussion or even to contestation that challenges them. It is interesting how, for example, Sen (2009) refused to lock in a final list of central capabilities and held that communities should decide through public reasoning which capabilities are important. The same principle suits futures work. Values should not always be brought ready-made to the table merely for finding the right ones or for balancing between them. The “value-ness” of the values themselves should be part of what is discussed. This also relates to participatory approaches: without deep value discussion, participation easily remains superficial: people do get to speak, but their words do not change the frame within which the speech is heard.

Values change

So far we have largely spoken as if futures were assessed with values that are relatively stable or acceptable to us. There is a certain practical sense in this. Without some assumption of the continuity of values, the assessment of futures would become nearly impossible. We could not say much about what a desirable (good and just, valuable) future is if we thought that all the criteria of assessment can change at any moment. Yet values are not unchanging.

A stark example of the nature of values is that only in 1833 did the British Parliament legislate the abolition of slavery. A generation earlier, many had regarded the slave trade as a natural, necessary, or at least acceptable part of the economic and social order. By the time of abolition this view had not merely changed but, in the eyes of many, turned morally shameful. Kwame Anthony Appiah (2010) has emphasized that the change was not driven primarily by new knowledge of the cruelty of slavery; that knowledge had been available earlier as well. The change was connected to moral sensitivity, to conceptions of honor, human dignity, and the civilized nation. We can know and still act wrongly if our action accords with wrong values; we can change our action by changing our values, not only by adding knowledge.

This is not a historical exception but an example of a broader phenomenon, one that is obvious but has for some reason remained hidden from futures research. Danaher (2021) has, in an eye-opening way, proposed axiological futurism, by which he means the systematic study of how human values can change in the future. The starting point is simple. When we look at the moral commitments of past societies, we find beliefs and practices that we now regard as deeply repugnant from the standpoint of morality and justice. The problem is that it is quite possible that future generations will look at some of our current ethical commitments with disapproval. Historical examples abound.

Cross-cultural variation also tells of the malleability of values. For example, Buddhist and Western moral communities can differ deeply in how they relate to the individual self, to desires, or, say, to righteous anger (see e.g. Flanagan 2017). Such differences do not mean that everything is acceptable as long as it is related correctly to a given culture, but they do show that the world of values is not simply given and universal. Values are connected to conceptions of the human being, community, suffering, responsibility, and the good life, and these conceptions are constructed in historical conditions.

Value change is driven by both intellectual and material factors. Intellectual drivers are new ideas, moral arguments, concepts, and modes of persuasion. Appiah's observation on moral revolutions is interesting here: changes often do not start from entirely new arguments, but from existing arguments beginning to connect differently to other things such as honor, esteem, identity, and shame (see Appiah 2010). Material drivers, in turn, relate to technology, the economy, energy use, and the physical conditions of life. For example, Morris (2015) has argued in his historical examination that the forms of energy use have shaped values concerning violence, equality, and justice. Verbeek (2011), in turn, has emphasized that technologies do not merely realize values but reshape them by opening and closing possibilities of action and by changing the concepts with which we understand the world. For example, it can be argued that the ultrasound device constitutes the fetus in new ways: as a patient, a person, and an object of choices.

Intellectual and material drivers do not, of course, operate separately. What we value guides what kinds of technologies we create and adopt, and these technologies in turn shape what we hold valuable. For example, the printing press did not merely spread existing ideas more widely, but created new kinds of publics, religious movements, scientific communities, and forms of political action. The internet did not merely speed up communication, but changed how we understand things like moral concern, attention, shame, solidarity, and hostility. This intertwining of the intellectual and material drivers makes futures research more difficult (which to focus on first?), but there is nothing we can do about this. Futures research is not supposed to be easy, especially when values are being discussed. Values are not an easy topic.

In a certain way one could of course think that the problem is not quite this large. Historians and anthropologists constantly study foreign worlds of value. They are able to reconstruct what it was like to live in a society in which the heavenly bodies were thought to orbit the Earth, honor determined social position, or gender roles were much starker than today. This possibility of reconstruction exists. But the study of past value worlds is retrospective. The researcher examines a frame that has already existed and of which traces remain: texts, institutions, practices, and concepts. In the case of a future foreign value world, such traces do not yet exist – at most weak signals. In addition, the present has a tendency to leak into our imaginings and to fill the alternatives with things we already know. This problem of presentism, that is, the influence of present-day conceptions, is a problem both in futures research and in thinking concerning the past.

For futures research a certain discomfort follows from this. Usually studies concerning the future are value-driven, because people care about the future precisely on account of their value commitments. Concern about climate change, existential risks, the erosion of democracy, or the control of artificial intelligence springs from the fact that we hold something important: human flourishing, justice, freedom, ecological sustainability, or the continuity of life. Yet futures research has all too rarely asked what will be valued in the future, what is desirable in the future. We study desirable futures and start from our own values. More rarely do we study what is desirable in the future – what values prevail. This is, putting it somewhat pointedly, odd.

Danaher's (2021) proposal of axiological futurism is more an opening than a finished theory. Its significance, however, lies in moving value change from the margins of futures research to being one of its objects of study.

This idea can be divided in two. First, one can examine what value change can be like. Not all change is equally deep, and not every change causes the same headache for the assessment of the desirability of futures. One must ask how difficult value change can make the assessment of the future. In this connection the discussion of transformative experiences and of the limits of imaginability is especially important. Second, one must face the normative puzzle that the recognition of value change generates: how can changing values be assessed without falling either into the arrogant domination of present values or into complete relativism?

These need not be thought of as separate problems, for good or ill. They relate to the same question of how futures research can assess futures if the grounds of assessment may themselves change.

When the experience does not yet exist

When speaking of the values of the future, the significance of experiences as shapers of values must be brought out. The philosopher L. A. Paul (2014) has described transformative experiences, by which she means experiences whose nature cannot be known in advance without undergoing them oneself, and which at the same time change the experiencer's central conceptions of the surrounding reality, their preferences, their aims, or their conception of themselves. One example is becoming a parent. Before the first child, a person can obtain information about parenthood, hear others' accounts, and form advance expectations. Still, no one can fully know what parenthood is like for them, nor how the experience will affect their values.

The original target of the critique was the idea of rational choice – the relevant information does not exist before the transformative experience. From this we quickly reach futures thinking: some future can be such that it cannot be properly assessed from the present, because the experience required for the assessment does not yet exist. What we lack is not merely, say, some statistic. Technologically modified humanity could be one example of a situation in which we attain transformative experiences that completely change our assessment of what is valuable and of what it is like to be in the future.

The idea of transformative experience can be extended; it is one limiting case. In transformative experiences the uncertainty is not the ordinary uncertainty about which future will come about. It is uncertainty about how the future would be experienceable and assessable if it came about. In futures research this is an important difference. It is an entirely different thing to say that we do not know whether something will happen than to say that we do not know how experience and valuation would change if it happened.

The claim can of course be questioned. It can be argued that corresponding, more familiar experiences give at least partial knowledge of what a transformative experience would be like. Others emphasize that the imagination can, as it were, be trained through narratives, art, simulations, and the experiences of others. This is important for futures work. Not all the limits of assessment depend in a deep way on whether we attain experiential understanding in some particular context. Some of the limits stem from the poverty of imagination, from the homogeneity of participants, or from the fact that sufficiently many kinds of alternatives have not been sought and generative thinking has not been stimulated. But some of the limits can still be deeper.

Between arrogance and relativism

If values can change, by what right do the people of the present assess the values of the future? And if the values of the future can deviate from present ones, must we accept any change whatsoever merely because in the future it may be held acceptable?

At first glance there appear to be two bad answers to this, and we face a dilemma. The first answer is evaluative arrogance. According to it, our present values are the right ones, and every future deviating from them is deficient insofar as it does not accord with those values, whatever future generations may say. Sometimes this attitude would seem to be right. It condemns futures that include, for example, slavery or arbitrary courts. The problem is that one cannot distinguish the futures that are appalling in respect of values from those that are merely different – there is no strict boundary. It is hard to come up with a sustainable argument for the claim that the values have already been found and may not change. A further problem is that, in the light of history, such an attitude of the superiority of one's values has also served as an instrument of cultural domination, when progress has been read exclusively in the terms of one culture. We do not want to do the same to the future.

The second (bad) answer is value relativism. According to it, every value system is as good as any other, and no trans-historical comparisons can be made. This may at first look even like a humble position. It would seem to respect the self-determination of future people and to leave room for values foreign to us. Humility can, however, conceal cowardice. If all value systems are held to be equally good, no future can be condemned, however horrifying it may be. For example, a future in which a technologically enhanced elite subjugates the rest of humanity and holds the subjugation morally justified would, from the relativist's perspective, be no worse than a future of universal flourishing, if the values of that future accept the arrangement.

This is the dark dilemma brought out by axiological futurism. The way out, however, is not to choose one of the extremes, but a method. In philosophy, the approach of so-called reflective equilibrium has been widely used. In it one proceeds by fitting together considered particular judgments and general principles, correcting each in the light of the other, until the whole is as coherent as possible (see Rawls 1971). We may, for example, notice that we have the principle “do not hurt”. Yet a police officer may have to use force to restrain a person, and this is right (a considered judgment). The principle changes into the form “do not hurt, unless a great threat to safety has arisen”. On the other hand, we may have the considered judgment “hurting is right in a boxing match”, but this conflicts with our principle. We ban boxing. In other words, in the method we add judgments and modify principles, back and forth, and the process does not come to an end anywhere. That is precisely why it suits the investigation and management of value changes; we make a change first over there, then over here, and gradually our value system and our conception of reality are renewed – but systematically and in a controlled way.

This gives humility against arrogance and backbone against relativism. Our present values are revisable, but revisability does not mean that anything goes.

It is fair to mention also a position according to which it is at least in principle possible that values should not change. Moral realism means the position according to which there exist objective moral truths that can in principle be discovered (see Shafer-Landau 2003). Then the endeavor to find the genuinely desirable futures is not arrogance but a striving for accuracy. But the realist, too, needs a method for forming and correcting their conception of which “values” really are values. Nothing prevents the use of reflective equilibrium. That is why reflective equilibrium serves both those who believe in objective values and those who take a more cautious attitude toward this.

For futures work a concrete instruction follows from this: within the value frame, a robust core and a fragile edge must be distinguished. Some commitments, such as the prohibition of arbitrary cruelty, respect for human beings, and some kind of fairness in the distribution of benefits and burdens, belong to the robust core. Other commitments are more fragile. They relate to more situation-specific applications, emphases, historical conditions, and culturally learned norms – different ways of conversing can be respectful (robust core) and still change (which would not cause a moral crisis). In futures research, too, there is perhaps reason to consider how the robust core and the fragile edge are presented when desirable futures are sought, and how deep into the core there is time and opportunity to go while the exercise still remains controlled.

Universal values reconsidered

What has been said above helps to sharpen the idea of universal values, about which, among others, Bell (2009 [1997]) wrote in his classic. Democracy, human dignity, equality, justice, and freedom are often understood as values that have, or should have, universal significance both now and in the future. Such talk is tempting and in part even necessary, but it easily turns into a heap of unspecified slogans if universality is understood as a ready-made content shared everywhere in the same way – without saying what exactly this content is.

One must therefore be more subtle. The search for universal values does not mean that all cultures would understand the values in the same way or give them the same content. It is perhaps more fruitful to examine what kinds of constraints any functioning morality must satisfy, and thereafter what kinds of different contents different cultures and different times give to these constraints. Wong's pluralistic relativism is, despite its name, a sharply delimited position, yet more fine-grained than empty universalism: in a certain sense there is more than one “true” morality, but not all moral systems are true. The core of the idea is that the task of morality is to regulate cooperation between people and to support the individual's inner order and flourishing. This task, together with the enduring features of human nature, sets universal constraints: a functioning morality must rest on reciprocity, offer those whom it binds reasons to accept it instead of mere coercion, and leave room, alongside the impersonal point of view, for personal ties (see Wong 2006). It is precisely such constraints that distinguish a defensible pluralism from boundless relativism.

It is important to study two things in parallel. On the one hand, one must ask what kinds of shared values people actually have. On the other hand, one must ask what kinds of shared values we could and should have. The latter is the more future-oriented question. Even if values valid nearly universally were found, they are not given ready-made or permanently valid. They change slowly, contestedly, and continuously, as a result of people's work, conflicts, institutions, and changing conditions. One cannot, however, jump straight to the values of the future and leave unasked what is already valued – and why.

What this means for practical futures work

What has been said above can be presented, from the standpoint of practical futures work, in the form of a few conclusions.

The first is that the assessment of scenarios and other images of the future should not be left to mere intuition. Most futures exercises assess their results informally: the team discusses, the client emphasizes certain things, participants find some alternatives interesting or worrying, problems are mentioned and answers are sketched. This is not wrong, but it is often insufficient. A clearer alternative is to assess futures with several explicit normative questions. We can ask whether people cross the threshold of sufficiency in a given future, what kinds of capabilities the future opens or closes, what kinds of relations it builds, and what forms of subjugation it may reinforce. This does not mean that every scenario exercise should be turned into a heavy philosophical analysis. Rather, the point is that the criteria of assessment are made visible.

The second conclusion concerns participation. The values used as yardsticks of desirability assessments should not always be defined in advance within a certain group of experts. Before futures are assessed, it would be worth reflecting together with those outside the expert group on what sufficiency means in the particular context, which capabilities are essential, and what forms of subjugation people regard as the most serious. Here the definition of value criteria is seen as part of futures work, not as preparatory work external to it carried out within a certain group.

Third, the futures described and their world of values must be in harmony with each other. There are no completely value-free scenarios whose desirability is, as it were, mentioned at the end. The values, institutions, technologies, and material conditions of a future come in the same package. For example, a scenario that assumes an enormous esteem for science and at the same time a complete collapse of research funding, without any explanatory mechanism, is internally problematic. Value and action do not meet, and the scenario is incoherent. Likewise, a scenario that assumes strong individual freedom and complete technological surveillance at the very least demands a meticulous explanation of how these fit together. Values do not float above the future; they attach to practices, institutions, and everyday possibilities.

Fourth, the assessor is part of the process, not outside it. Futures work affects action, and action affects the future. When some future is described as desirable, this description can steer resources, attention, policy, and imagination. Therefore the describing of the values of a future and the recommending of them are not fully separable from each other. A value judgment concerning the future participates, often modestly to be sure, in producing the conditions in which that judgment is later either confirmed, changed, or overturned.

Fifth, the limits of the assessment of desirability must be remembered. Some futures can be so thoroughly different that we simply do not know what it would be like to live in them. And if we do not know what it would be like to live in that future, we are not able to assess very strongly the valuations connected with it.

Finally, it must be remembered that values change. We should study more the mechanisms concerning value change and the possible futures of values. In addition, we should answer the question of whether we accept the change or not – whether it is morally permissible. From this a dilemma arises for us between present and future values.

None of this makes normative futures work – that is, the search for and pursuit of desirable futures – impossible. Rather, it makes this work continuously unfinished in a recognizable and intelligible way. We do not wander on the basis of intuitions but constantly reflect on why we value something and why our values can change. This is actually an important strength. Futures research that recognizes the historicity and limitedness of its own yardsticks is in a better position than futures research that imagines the permanence of its own yardsticks. Let us value change. And let us change our values – but sensibly.

References

• Appiah, K. A. (2010). The honor code: How moral revolutions happen. W. W. Norton.
• Aristotle. Nicomachean Ethics. (Finnish translation Nikomakhoksen etiikka by Simo Knuuttila. Gaudeamus.)
• Bell, W. (2009). Foundations of futures studies: Human science for a new era. Transaction Publishers. (See especially volume 2 of the two volumes.)
• Bentham, J. (1789). An introduction to the principles of morals and legislation. T. Payne and Son.
• Berlin, I. (1958). Two Concepts of Liberty. Oxford: Clarendon Press.
• Casal, P. (2007). Why sufficiency is not enough. Ethics, 117(2), 296–326.
• Chang, R. (Ed.). (1997). Incommensurability, incomparability, and practical reason. Harvard University Press.
• Danaher, J. (2021). Axiological futurism: The systematic study of the future of values. Futures, 132, Article 102780.
• Flanagan, O. (2017). The geography of morals: Varieties of moral possibility. Oxford University Press.
• Frankfurt, H. (1987). Equality as a moral ideal. Ethics, 98(1), 21–43.
• Fricker, M. (2007). Epistemic injustice: Power and the ethics of knowing. Oxford University Press.
• Inayatullah, S. (1998). Causal layered analysis: Poststructuralism as method. Futures, 30(8), 815–829.
• Kant, I. (1785). Grundlegung zur Metaphysik der Sitten. Johann Friedrich Hartknoch.
• Mill, J. S. (1863). Utilitarianism. Parker, Son, and Bourn.
• Morris, I. (2015). Foragers, farmers, and fossil fuels: How human values evolve. Princeton University Press.
• Nussbaum, M. C. (2000). Women and human development: The capabilities approach. Cambridge University Press.
• Paul, L. A. (2014). Transformative experience. Oxford University Press.
• Rawls, J. (1971). A theory of justice. Harvard University Press.
• Scanlon, T. M. (1998). What we owe to each other. Harvard University Press.
• Shafer-Landau, R. (2003). Moral realism: A defence. Oxford University Press.
• Sen, A. (1992). Inequality reexamined. Clarendon Press.
• Sen, A. (2009). The idea of justice. Allen Lane.
• Verbeek, P.-P. (2011). Moralizing technology: Understanding and designing the morality of things. University of Chicago Press.
• Wong, D. B. (2006). Natural moralities: A defense of pluralistic relativism. Oxford University Press.

[AI assisted translation from Finnish non-AI text.]

Key Points

• Without a conception of time there is no conception of change — and no futures thinking. Time is not a uniform background within which things happen, but a central part of how the future is thought about, studied, and built.
• The same moment contains many temporalities. In a hospital, the patient’s time is waiting, the physician’s time is urgency classifications, and the administration’s time is budget periods — and the future appears differently to different actors precisely because of these differences.
• The choice of time horizon is an exercise of power. It delimits which futures enter the discussion: an organization looking at the quarter sees a different world than one looking toward 2050.
• Not everything significant happens in long arcs. Climate policy needs decades, crisis management hours, and algorithmic decision-making seconds — the right question is which time horizon makes a phenomenon intelligible.
• The choice of method is at the same time a temporal choice. Every method has its own time-needs, and one must ask not only what kind of time a method requires but also what kind of time it produces.

Why must we talk about time?

Time seems simple in everyday life; it passes at a certain pace, whether we are in a hurry or not, whether it is as we hoped or not — we will return to this. Normally, however, we merely look at the clock to know what time it is now, open the calendar, schedule meetings and activities at points in time, expect an answer within a certain time, perhaps even miss a train or make it just in time.

Below, we treat time both from the standpoint of the human being — the individual, the organization, the collective, and so on — and as the research framework of futures research. The choice is deliberate: time must be understood more broadly in order to understand time as a research framework for the future. These go hand in hand.

In futures research, and in thinking about the future generally, time is not merely a uniform and unnoticed background within which things happen. It is an absolutely central part of how the future is thought about, studied, and how the future can be influenced. When we speak about the future, we always speak about time as well — the future is, of course, itself a temporal concept. There are, however, particular reasons to reflect on how far in time we look, at what speed change happens, from whose perspective time and change feel long or short, and when the right moment to act is, for whom.

Futures research does not mean certain prediction, as we have noticed (see the text on the conception of knowledge). The task of futures research is rather to examine possible, probable, and desirable futures, and how, among other things, present-day decisions, expectations, technologies, and societal structures shape what is to come (Bell, 2009; Inayatullah, 1993). Time is present here all the time, to put it somewhat ironically. Without a conception of time there is no conception of change. Without an idea of change there is no futures thinking (Bell, 2009). In futures research, time is thus connected above all to change. If nothing changed, the future would be only an extension of the present. The future is interesting precisely because things can go otherwise.

That is why the question of time in futures research, and in the life that is its object, is fundamental. It is not enough to ask what can happen in the future. We must also ask: on what timescale, in whose time, at what and whose rhythm, and with what consequences? Things change, and they change in someone’s time and in some time.

Marcus Bussey has condensed this idea by pointing out that futures researchers work with time and within time: they operate amid different temporal cultures and rhythms, not “outside” them (Bussey, 2017). A futures researcher does not merely choose a year, such as 2035 or 2050, and place things to happen before or around it. Whoever studies and reflects on the future must ask what kind of temporal assumptions make some futures credible and others difficult to think. It is improbable that a car built in 2006 will still be running in 2050; it is probable that an apartment building constructed in 2016 will still be standing in 2046 — excluding catastrophes, which unfortunately do occur. It is possible that in 2030 the football World Cup will be won by a different country than in 2022. These are only some examples.

Timing is thus a central concept when reflecting on possibilities (from far-fetched to highly credible). The same thing can be too early, too late, or precisely well-timed depending on the actor. A technology can exist before society is ready to use it responsibly. A political decision concerning a crisis can be effective only if it is made before the crisis escalates. A vaccine can be completed before a pandemic is out of control — that is, it can be; nothing guarantees this.

In futures research, then, time is not merely duration or some “container” in which events are located. Attention must be paid to it as if it were an actor in itself, to put it somewhat pointedly. Time in the hands of an actor is a tangle of issues to which we return below and which has significant features.

Conceptions of time and temporalities

Time is often thought of linearly, as a straightforward, uniform continuum. The past is behind, the present is here, and the future is ahead. As time passes, the present and eventually the future become part of the past. This is a familiar way of thinking especially in modern society, where development, growth, planning, and progress have been important matters, also from the standpoint of narrative, governance, and experience. (See e.g. Bell, 2009; Thompson, 1967.)

Linear time undeniably suits many questions of futures research well, even though we can call it into question beneath the surface. If we want to consider, for example, the transformation of energy production by 2050, the development of the education system over the next twenty years, or the effects of technology on working life, we need a conception of time that moves forward. Only such a time makes it possible to arrange steps into a path toward the future and to examine whether the right step has been reached at the right time in the right context. If we want transport to be electric by 2050, in 2046 it may have to be concluded that the path did not follow the desired timing.

But time is not only linear. Or rather: linear time is not quite what it appears to be.

Many things are built on rhythms and repetitions. There are many examples in different contexts. Seasons, days, and hours repeat. Likewise, for instance, the rhythms of care work, economic cycles, electoral terms, and generations create cycles within linear time — no work shift is identical to another, yet they repeat all the same. One can adopt a kind of cyclical conception of time without assuming that nothing changes. Rather, change then means that something is achieved (or an end result is arrived at) through repeating phases, rhythms, and recurrences. In sport this is familiar. But also, regrettably, for example in warfare or in preparing for it.

In addition, experienced time can always be distinguished from linear (perhaps cyclically forming) “objective” time. It does not always run according to the clock or other external recurrence. “Time is long for the one who waits.” For the busy, time passes quickly. In uncertainty the future can feel heavy, and in an inspiring situation it can open up as possibilities. In futures research this experienced time matters greatly, because people do not react only to facts and forecasts. They also react to how the future looks in their life situation or as a certain kind of actor. Whatever the researchers say, it has no effect and is of no interest if a person lives outside the researchers’ discourse where future prospects are concerned.

We can therefore emphasize the concept of temporality. Temporality means how time is constructed, experienced, and ordered in different situations, systems, and relations. The difference between time and temporality is central: Time is. Temporality makes it part of life. In the same time, in the same situation, there can be many temporalities. Consider, for example, a hospital. For the patient, time can be waiting, pain, or uncertainty. For the physician, time can be appointment slots, urgency classifications (a temporal concept), and care processes. For the administration, time is connected to the use of resources, work shifts, and budget periods. For the technological system, different situations and actors can be timestamps, response times, and data updates.

All these temporalities are real. Their length, their experiencer, and their effects can be very different indeed, as our example shows.

In futures research, recognizing multiple temporalities helps us understand why the “same” future can appear differently to different actors. We say that something will perhaps happen at some moment in the future. One then sees an opportunity, another a risk, a third a delay, and a fourth a change that is too fast. Time is thus not only a question of measures (of measuring time) but also a question of interpretation, experience, and power (see e.g. Bussey, 2017; Virmajoki & Laakkonen, 2024b).

People and communities also orient themselves toward time in different ways. This can be called time orientation. In futures research this means that thinking oriented toward the past seeks continuity, experiences, traditions, and the lessons of earlier solutions. Thinking oriented toward the present concentrates on the situation at hand: what is urgent, what must be done now, what the current situation is — and there can be different views on this; the present is always contested. Thinking oriented toward the future, in turn, asks what can happen, what is worth pursuing, and what consequences current choices may have. Let us be precise: future-oriented thinking is not the same as the time orientation of futures research. The latter can contain many different time orientations. We learn from the past, reflect on the present, try to grasp what is to come.

An organization, too, can be time-oriented in different ways simultaneously. It can speak in its strategy about the year 2046 but in practice operate on the terms of the next budget year. It can emphasize renewal but value personnel on the basis of the past, whether this is right or not. It can hold a crisis meeting in the middle of the day. Such — in a certain sense contradictory — time orientations are interesting for futures research because they affect which futures become credible or possible from the standpoint of action (see e.g. Brier, 2005; Nordlund, 2012).

Time orientation is also visible in emotions — of an individual or a collective. If the past is seen as a lost golden age, the future can feel threatening. If the present is experienced as a crisis, the future can narrow into a question of survival. If the present is seen as resource-rich and open, the coming change can feel mostly like a decision about where to direct activity. Emotions and agency are bound together through temporality.

Time horizons from long to short in futures research

Timespan or time horizon means how far into the future the examination of a topic or question is extended. In futures research, common time horizons are, for example, five, ten, fifteen, twenty, or fifty years (Brier, 2005; Nordlund, 2012).

Such delimitations are not arbitrary. If we set the question about the future too close, it is difficult to reflect on (how could anything have time to change?), but if the question is too far away, the reflection can become unclear, even sci-fi-like. How do we know what all has changed around the matter if the time horizon is 50 years? Who knew what the year 2026 might contain, even though Apple released a certain computer in 1976?

Time horizons are not, however, freely choosable by the researcher. Often time horizons are so essentially tied to the practices of planning (such as government terms, strategy periods, investments, technological development, or generation-length changes) that the time horizon must be set on the terms of the user of futures knowledge. For the sake of integrity, however, empty promises cannot be made: not if a given time horizon is impossible from the standpoint of the realities of research — for example, the assembling of a research team and the collection of data.

Here it must be taken into account that the choice of time horizon changes what can be seen of the future. If we look only at the next year, slow developments can remain hidden. Climate change, population ageing, changes in the level of education, or the decay of infrastructure do not necessarily look dramatic in a short examination. If, on the other hand, we look only very far, the contradictions of the present, the pressures of decision-making, and people’s everyday experiences can disappear from view. We do not want to lose our grip on either, or give up on either.

The choice of time horizon in futures research and futures thinking is therefore also an exercise of power. It delimits which futures enter the discussion and which remain outside it (Bauer, 2018). When an organization chooses only the next quarter as its interval of examination, it sees a different kind of world than an organization that asks what its activity means in 2050. Inside these organizations, different things are discussed.

Futures research is often associated with relatively long time intervals, as we saw above. This is understandable. Many questions important for the future require examination reaching years or decades ahead: energy production, urban structure, education, population development, climate, democracy, and technological disruptions.

Yet the future is not built — nor present and experienceable — only in long arcs.

Short moments can have great significance. A decision, a delay, a misinterpretation, the speed of reaction, or a choice steered by a user interface can affect almost immediately how the totality of possibilities is experienced. Research has shown how, especially in digital and algorithmic systems, the meaningful time intervals can be minutes, seconds, or even shorter timespans (Virmajoki & Laakkonen, 2024a).

The impact of long and short time intervals on experiences and knowledge concerning the future shows how different questions need different time horizons and how we cannot avoid the differences between time horizons even in futures research. Climate policy needs assessments reaching decades ahead. Crisis management needs assessment reaching hours ahead. A transport system often needs minutes. Algorithmic decision-making may need seconds or less.

What is important in futures research is to ask which time horizon makes the phenomenon intelligible. There is no single correct time horizon to which futures research is inevitably attached.

Time-needs and the methods of futures research

Time-need (aikatarve) is a useful concept when one wants to understand why time is a different frame — and a different, even almost concrete, thing — for every actor, however strange this may sound at first. Time as something concrete? One must begin from the fact that questions concerning time are not exhausted by duration, that is, how much time some task takes. The concept of time-need (Virmajoki & Laakkonen, 2024b) describes how people, organizations, and technological systems

• need,
• measure,
• represent,
• use, and
• control time.

The first aspect of time-need is duration. How long does some process take? A forest grows slowly. Trust is built slowly. A message travels fast. A stock price can change in an instant. An algorithm decides in a second where the elevator goes.

The second aspect is measurement. In one situation a rough estimate suffices: the change will happen sometime during the next decade. In another situation a suitable day for a meeting can be agreed on — “let us agree on the exact time later.” In intensive care or in software systems, seconds can matter.

The third component is the representation of time. Time is depicted with calendars, timetables, and clocks, but — especially in futures research — also with forecasts, scenarios, and visualizations extending into the future. These do not merely depict time. They also steer how people understand time, the progression of the future, and their own possibilities for action within this course.

Time-need also includes timing, coordination, control, and experience. The same act can be right at one point in time and wrong at another. The rhythms of different actors must be fitted together. Someone sets the deadlines. Someone waits; another makes others wait. For example, it must be ensured by means of stops that trains do not collide (coordination and timing). A delay can be communicated to the passenger in many different ways — or sometimes not at all, or in a completely misleading way (control — that is, control of others’ time). Waiting in a train without air conditioning can feel endless (experience). (See more on the concept of time, e.g., Virmajoki & Laakkonen, 2024b.)

The methods of futures research also relate to time in different ways — they have different time-needs (Virmajoki & Laakkonen, 2024b). A forecast often seeks to estimate what will probably happen within a certain, relatively short time interval. A set of scenarios, in turn, describes alternative courses of development: what could happen if some things change. In the Delphi method, experts’ views are collected and refined in stages — often according to who has time to answer and when, and who analyzes the answers and when.

From this it follows that the choice of method is at the same time a temporal choice. When one commits to a method, one commits to examining things that (possibly) happen after a certain time, and to a research project of a certain duration. For example, scenario work requires time for discussion, interpretation, and the construction of alternatives. The examination of trends can sometimes happen faster and even in an automated way, but it can also narrow the gaze to changes that are already visible. Long-term foresight often needs broad participation and commitment, because decisions concerning the future touch many actors (see further, e.g., Bell, 2009; Bussey, 2017; Virmajoki & Laakkonen, 2024b).

Therefore, in futures research it is not only a question of which method is used. One must also ask what kind of time the method needs and what kind of time it produces.

Time, knowledge, narratives, and power

Knowledge concerning the future is always temporal, as we saw above. We know something about past developments, something about the structures of the present, and something about possible directions. But the further we look, the more uncertainty generally grows.

This does not make futures research or futures thinking useless. On the contrary — because uncertainty is one reason why alternative futures are worth studying. If the future were ready-made and knowable with certainty, futures research would not be needed in the way it is now. The value of futures research generally lies in helping to identify alternatives, assumptions, consequences, and points of choice (Bell, 2009; Inayatullah, 1993), although we have also gone through exceptions.

Time also affects what kind of knowledge is considered credible — time is an epistemic factor, a factor of knowing. It affects our conceptions of what we know and why. In the short term, a statistical forecast can look strong because it is based on previous, solid statistics. In the long term, discontinuities, surprises, and societal disruptions can become emphasized — things of which, for example, weak signals may tell. On the other hand, in a very short time, people’s interpretations, reactions, and misunderstandings — the not-having-time-to-know — can steer the course of events in ways that a long time horizon easily smooths out of sight. (Virmajoki & Laakkonen, 2024a.)

Knowledge of the future is thus not only a question of what is known. It is also a question of in what time it is known. It is interesting to note how stories are temporally constructed presentations that also affect our conceptions of what is to come.

In futures research, narratives often matter, and they must at least be taken into account. People make sense of the future with the help of stories: will development continue in the current direction, is a collapse ahead, will a great transformation occur, or will a new equilibrium be found? (See, e.g., Dator, 2009, on such archetypes contained in our narratives.) Narratives are always temporal. They have beginnings, turns, continuities, ruptures, and possible endings. That is why images of the future are not merely lists of events but temporal orders in which some things appear to follow from others.

Futures researchers work with temporal models and narratives: they can make visible what kinds of conceptions of time delimit the imagination of people and organizations (Bussey, 2017). This does not mean that all “stories” are equally good or that facts do not matter. The point is that knowledge and interpretations concerning the future are often organized into some kind of temporal narrative.

That is why in futures research one can ask what the dominant story about the future is, or who tells it. One can ask what the existence or the telling of the story makes possible, or what it excludes.

From this we arrive at the question of the relation between time and power, which we touched on above in connection with time-need (the aspect of control).

Time has a history — that is, the way we commonly conceive of time has one. In the history of industrialization, clock time, working hours, and the rhythm of production changed people’s everyday life and work fundamentally. It can be said that time began to be seen as a resource that could be used efficiently or inefficiently (Thompson, 1967). In contemporary society the same question repeats itself over and over. In the nature of work this is clearly visible. Project work, the platform economy, constant reachability, real-time monitoring, and digital services are becoming emphasized. With each of these we can ask whose time is saved, whose time is fragmented, who has to wait, or whose future is relevant to strategy in the long run.

These are questions of futures research, not merely facts of everyday life into which time is entwined ever more tightly. If decision-making examines only short-term efficiency, it can shift costs to future generations. If, on the other hand, only the distant future is spoken of, the inequality of the present can be left aside. It must be repeated that the time horizon chosen as the object of examination is not an innocent choice (see, e.g., Bauer, 2018; Milojević, 2008; Virmajoki & Laakkonen, 2024b).

The past, the present, and responsibility in time

In futures research, the past is not merely time left behind. It is material, memories, experience, and narratives with whose help the future can be made intelligible. The past can open up possibilities; it can (and should), for example, remind us that things have changed before. It can show that the present that is taken for granted has come about through choices, conflicts, and contingencies. We can perhaps influence these consequences by understanding their causal history.

However, the past can also exclude possibilities or render them invisible. If some solution has once failed, it may be regarded as impossible in new circumstances as well — perhaps we gave up easily. If the past is told as only one inevitable story, alternative futures can narrow. So-called critical futures thinking therefore pays particular attention to how the past is used. According to Inayatullah (1993), the many meanings of time must be capable of being worked with, not merely taken as given. Time, because it does not “just be,” is a matter that must be examined critically. Images of the future are always constructed in relation to how the past and the present are understood. This article has sought to give a rough picture of the matter.

It must also be noted that the present is not merely a thin line between the past and the future, a mere blink of an eye in the stream of time. It is always a kind of crossroads. In the present, the consequences of past decisions, current structures, and expectations concerning what is to come all meet. Decisions are made in the present, but their justifications often come from different times: “this is how it has always been done,” “such a change is being planned,” “we must prepare for this better than before.”

From the perspective of futures research, the present is interesting — somewhat paradoxically — precisely because it is not finished, or at least not unambiguous and experienced in the same way by all. The present contains alternatives — on what else than the present could alternatives be based? Not all possible futures are equally probable, and not all can be chosen, but the present, if anything, is what we can influence and through which we can build the future.

Many decisions of the present affect far beyond the decision-makers’ own lives, and further in time. Climate change, the weakening of biodiversity, infrastructure, the societal effects of artificial intelligence, and the final disposal of nuclear waste are examples of questions in which our responsibility in the present extends far into the future. Future generations cannot participate in today’s decisions, but today’s decisions affect their lives and their possibilities.

For this reason, the choice of time horizon is also an ethical question. If we look only nearby, we can shift costs and burdens onto those who are not there looking with us. If we look further, we can perceive responsibilities that otherwise might not fit within the scope of decision-making (Bussey, 2017).

Responsibility in time does not, however, mean only responsibility for the distant future. It also means taking into account the different time experiences of people in the present and weighing their ethical position now and in the near future. Slow crises, fast crises, the rush of everyday life, prolonged uncertainty, and waiting are all temporal experiences that matter to the ethical side of the construction of futures.

In conclusion: time is an instrument of questioning

In futures research, time is not merely a year in the title of a report, such as “Finland 2050.” Time, temporality, time-needs, power in time, and so forth are ways of understanding reality at a deep level. Futures research is in a privileged position if it genuinely takes into account the understanding it has itself constructed of time and of the matters related to it, from duration all the way to ethics.

Therefore, in futures research one must ask what changes fast and what changes slowly; on what time horizon a phenomenon becomes visible; whose time defines the conception of reality and its “timeliness” — when is the right moment to act; and what kinds of interpretations of the past delimit the future. If one wants to think about the future seriously, one must ask these questions. This concerns each of us, including those who practice futures research professionally.

Finally, there remains the question of what goes unseen if we choose the wrong time, the wrong measure, the wrong determiner of temporality, and so forth. We do not want to be blinded because we stepped into the wrong time.

When a person encounters futures research for the first time, it must be grasped that the future is not merely something that happens later. It is a continuously constructed temporal relation between the past, the present, and possible future worlds.

References

Bauer, A. (2018). When is the future? Temporal ordering in anticipatory policy advice. Futures, 101, 36–45.

Bell, W. (2009). Foundations of futures studies: Human science for a new era (Vol. 1). Transaction Publishers. (Original work published 1997)

Brier, D. J. (2005). Marking the future: A review of time horizons. Futures, 37(8), 833–848.

Bussey, M. (2017). Time’s calling: Time, timing, and transformation in futures work. World Futures Review, 9(4), 236–247.

Dator, J. (2009). Alternative futures at the Manoa School. Journal of Futures Studies, 14(2), 1–18.

Inayatullah, S. (1993). From “who am I?” to “when am I?”: Framing the shape and time of the future. Futures, 25(3), 235–253.

Milojević, I. (2008). Timing feminism, feminising time. Futures, 40(4), 329–345.

Nordlund, G. (2012). Time-scales in futures research and forecasting. Futures, 44(4), 408–414.

Thompson, E. P. (1967). Time, work-discipline, and industrial capitalism. Past & Present, 38(1), 56–97.

Virmajoki, V., & Laakkonen, M.-P. (2024a). Don’t push the wrong button: The concept of microperspective in futures research. Futures & Foresight Science, 6, e183.

Virmajoki, V., & Laakkonen, M.-P. (2024b). Time and futures: Analysis of time-needs in futures research. Futures & Foresight Science, 6, e190.

Key points

• Justified knowledge can be produced about the future, even though it cannot be observed. Not all claims concerning the future stand on the same line — some conceal uncertainty, some make it understandable.
• Knowledge is not merely true sentences. A person can know that the population is ageing, yet still understand poorly what it means — and it is precisely understanding, not a list of claims, that is the aim of futures studies.
• The object of futures studies is the space of possibilities. The field does not ask what will happen, but what kinds of futures are possible on the basis of present conditions, structures and values — and how these conditions can be changed.
• The choice of method is always an epistemic choice. Every method, from trend analysis to layered analysis, contains assumptions about who knows, where knowledge is located, and why the present tells us something about what is to come.
• Good futures knowledge tells what quality it is of. A scenario should not be read as a forecast, nor a vision as a neutral description — good research says openly when it anticipates, when it assesses, when it aspires and when it warns.

Introduction: What is being spoken of when we speak of the future?

Everyone continually makes decisions on the strength of their conceptions concerning the future.

A student chooses a field. A municipality plans and alters its school network. A city makes decisions about transport. A company invests in technology. A state does, or refrains from doing, climate policy. In all these situations one has to take a stand on, or to steer purpose and ends towards, something that does not yet exist and that cannot yet be observed. One has to take a stand on, or commit to, what kind of future is possible. We ask: What should be aimed at? What should be avoided? What kinds of consequences can the decisions of the present moment have?

The point of departure of futures studies, put in this way, is simple, but it opens up problems relating to knowledge: the future cannot be observed and it can be influenced; how, then, can anything be said about it? Or, more precisely:

How can anything be known about the future?

Let us begin unpacking the problem from the other direction. Of the past there are traces, among them documents, statistics, memories (and the documentation of these) and objects. The present moment can be observed, measured, tested and otherwise examined. To be sure, even though there are sometimes enormous disagreements, this can in principle be done. Of the future, however, there is no similar observational material, because the future has not yet happened. From this it does not follow, however, that speaking about the future would be mere "guessing" or "speculation." Not all claims and thoughts concerning the future stand on the same line as regards their credibility.

Some are better justified than others, and often these better-justified claims and thoughts help in dealing not only with the future as such but also in understanding the present moment and the possible courses of development towards different futures.

Some are more poorly justified. These thoughts and claims conceal uncertainty, mix hopes with forecasts, or give an impression of certainty where there is no certainty.

The examination of the conception of knowledge in futures studies should begin from this difference, and from the tension between the past–present axis and the future. The future is not known exactly in advance, but justified knowledge can nonetheless be produced about the future — this is the point of departure of futures studies. Knowledge about the future is knowledge of

• Possibilities,
• Conditions and conditionalities,
• Relations of dependence,
• Uncertainties,
• Values, and
• The consequences of present choices.

Futures studies does not, then, aim to look at the future as though at a finished picture. Rather, it studies what kinds of futures can be justifiably discerned from the present moment. This can be called the study of the space of possibilities: the future is examined as a totality of different possible, credible, probable, desirable and to-be-avoided courses of development and outcomes (see e.g. Bell, 2009; Virmajoki, 2026b).

Knowledge is not merely true sentences

When we speak of knowledge, we easily think of individual true claims that we believe for one reason or another — because of some ground. We know the prevalence of some matter because the statistics on the phenomenon are of a certain kind. We know that some decision was made on a certain day because it has been so recorded. We know that some natural-scientific phenomenon works in a certain way under certain circumstances because we have observational material. We know that a medicine has an effect because it has been studied with a randomised trial.

Scientific knowledge is not, however, merely a collection of true sentences and the grounds that serve as their continuation-sentences.

Science also strives to explain, to model and to understand. An explanation answers the question of why or how something happens. Often this requires a model that describes, at least roughly, relations of dependence. Understanding, in turn, means the capacity to see how things relate to one another, and the capacity to command explanations and to act on the strength of them. Nowadays it can also be noted that understanding is something more than merely knowing a true claim or receiving an explanation. (See e.g. de Regt, 2017.)

Understanding is the capacity to use a theory to articulate, explain and anticipate phenomena. What understanding requires in each case depends on the historical, social and disciplinary context: a theory is understandable when its conceptual tools and the researcher's skills make its fruitful use possible. These differ from one another at different times and in different fields. This is central in futures studies: for example, a person can know that the population is ageing, but still understand poorly what this means for the future of society. Attaining understanding in this connection requires asking, for example, how ageing relates to the labour market, to care, to public finances, to housing, to technology, to migration and to justice between generations. It is not a matter merely of piling up claims concerning facts.

This difference is very significant. A great many individual facts can be known, but the overall picture can still be weak. Futures studies needs precisely the overall picture, because questions concerning the future do not generally focus on an individual feature of the future. As was set out above, they relate, among other things, to systems, choices, values and uncertainties.

In the philosophy of science a few basic concepts can be distinguished from one another:

• Observation means that which can be measured, seen, recorded or otherwise verified.
• Theory articulates comprehensively how things relate to one another, sometimes at a very abstract level and producing basic concepts.
• Model is a more simplified description or representation of some phenomenon than the theory behind the model (if there is one).
• Explanation tells why or how something happens.
• Understanding means the capacity to discern connections and consequences, and the capacity to assimilate explanations.
• Justification tells why some claim or assessment can be held acceptable.

These are not detached from one another, but there is reason to distinguish them. To understand the epistemic nature of futures studies all of them are needed. We need observations of the present moment: what kinds of trends, decisions, structures and changes are already visible? We need theories, so that we understand, so that we can lean on broader conceptions of the nature of reality — whether it be a matter of nature or of the human being (insofar as the line-drawing is nowadays meaningful). We need models, because complex reality cannot be dealt with as such or all at once. We need explanations, so that it can be discerned why something has happened and thereby why some future can happen. We need understanding, so that we can make use of the other epistemic elements when we examine the future. Justifications are needed so that claims concerning the future can be distinguished into ones of differing quality.

All this may sound like abstract hair-splitting, but the thought is not only necessary to grasp but ultimately also simple. The value of knowledge concerning the future does not lie only in whether some claim later hits the mark. The value of that knowledge lies in its helping us to see and to grasp what kind of situation we are in now, and what kinds of courses of development, contingencies and choices could change it — and change it into a particular future.

Why is the future a special object of study?

The future is a special object of study because it does not yet exist as an observable totality of events. This distinguishes the future from the present and from the past, at least in part. For the future is not detached from the present moment either.

In the present moment there are already things that make some futures more probable, or rather (so that we may avoid an excessive quantitative association) more possible than others. The demographic structure, the boundary conditions of nature, technological know-how and the degree of development of technologies, infrastructure, institutions, economic structures, cultural conceptions and political decisions all affect what kinds of futures we are moving towards and which ones we drift away from.

We can take a few examples, in all of which a certain relation of dependence drives us towards a future and we can say that future to be possible — even more possible than others. If a city is built on the basis of private-car traffic, this affects where shopping destinations are located and what kinds of emissions arise in the city in the future. If a country invests heavily in a certain energy system, this affects what kinds of ties the country must have to the availability of certain raw materials and what technology the country must have. If an education system is based on a certain conception of competence, this affects what kinds of experts there will be in society, or in a part of it, in the future.

The future, then, is not known, but the present moment is anything but a matter of indifference for the future.

This is the basic idea of futures studies. The field studies the relation between the present moment and possible futures. Futures studies does not ask what will happen, but what kinds of futures are possible on the basis of present conditions, actors, structures and values. At the same time it can ask how these conditions can be changed. The epistemic task of futures studies relates precisely to the fact that the future is not an object of study existing as such, but a space of possibilities opening up from the present moment (see e.g. Virmajoki, 2026b). Knowledge about the future is knowledge of the structure of the space of possibilities. In other words, knowledge about the future is a conception, based on the present and the past, of what can happen.

At this point the reader may ask: if the future has not yet happened, is not all knowledge concerning it uncertain; and is the knowledge not merely "knowledge"?

Once we notice that futures studies does not put forward, brazenly, claims about individual future situations, uncertainty does not mean ignorance. A doctor, too, does not always know for certain how an illness will progress. Still, they can produce a justified assessment on the basis of risks, symptoms, research knowledge and treatment options. In addition to this, the patient's action can affect how the ailment progresses. (This is not, unfortunately, always possible; but just as there are incurable illnesses, so, for example, the destruction of the climate cannot always be undone after a certain point.) In the same way, futures studies does not remove uncertainty, but strives to make it more understandable — to tell why the future is uncertain: what affects it and what relevant knowledge we lack.

Forecasting, explaining and understanding

A common misconception is that futures studies would be the same thing as forecasting. A forecast is ultimately only one form of knowledge concerning the future. In the by now canonical literature of futures studies it has indeed been emphasised that the study of futures does not reduce to the forecasting of one future course of events, but relates also to alternatives, choices and values (see e.g. Bell, 2009 [1997]). The matter must nonetheless be gone through; the mere claim of "one form of knowledge" does not yet tell us anything.

No one disputes the significance of forecasts. A forecast answers the question: what will probably happen?

Forecasts are needed, for example, in assessments of demographic development, the economy, weather phenomena or emissions development. They are useful especially when their grounds can be made sufficiently clear and when the phenomenon under study behaves in a way that can be modelled. A forecast that is not based on these is questionable; perhaps guess would be a better term in that case.

In research concerning the future it is worth distinguishing at least three epistemic aims, which we dealt with above:

• Forecasting asks what will probably happen.
• Explaining asks why something happens or could happen.
• Understanding asks how things relate to one another and how different courses of development can arise.
• Understanding of possibilities asks how understanding, explanations and forecasts can be joined together, so that a bundle of different representations concerning the future can be created, whether these be visions, scenarios or any set of claims whatever that describe the future.

In futures studies, then, an explanation answers a different kind of question than a forecast. If, for example, a scenario is drawn up of how an energy crisis could arise, the value of the scenario does not necessarily lie in its forecasting the crisis, but in its explaining what kinds of dependencies among, for instance, decisions, markets, infrastructures and political tensions could together lead to a crisis. Such an explanation is conditional, that is, it sets out conditions behind the phenomenon or for its emergence: if certain things happen and certain mechanisms operate, a certain kind of future can happen.

Conditional explaining can be understood as "what if things were (had been) otherwise" thinking. If some factor changes, what would follow from this for some other matter? If, for example, some decision is made, what kinds of chains of effect can it set in motion? Such thinking relates to the discerning of causal dependencies and does not require that we claim a certain future to be realised for certain (on causal explaining see e.g. Woodward, 2003).

Understanding is an even broader aim, the attainment of which the knowledge of futures studies tries to offer. As we saw, understanding is the capacity to discern connections and consequences, and the capacity to assimilate explanations. Understanding can indeed be thought of as a skill. No knowledge, even were we to receive from an oracle pictures of the future, suffices, unless we are able to attain understanding. An important aim of futures studies is indeed the increasing of understanding without claiming that the future is predictable — futures studies can, with the knowledge it produces, support the use and the role in life of a certain skill: understanding. For this reason good futures studies does not always remove uncertainty. Uncertainty helps to understand, if understanding is a skill: uncertainty tells where one must be careful. On the other hand, futures studies can instead help to identify what is relatively stable or what may change surprisingly. In this way futures studies can be thought of as research that produces understanding, not as a mere forecasting machine (on understanding see more closely e.g. de Regt, 2017).

The space of possibilities

The central object of study of futures studies can be described as the space of possibilities. By this is meant the totality of those different futures that can, in some way, be justifiably sketched out from the present moment (Virmajoki, 2026b). These futures are, to be sure, always the futures of some subject-matter, such as the futures of security policy or of working life. No one can discern the future in its entirety. Knowledge concerning the future is thus knowledge of spaces of possibilities.

The space of possibilities of some subject-matter is not, however, a mere list of alternatives. It is rather a map and a space in which we can already look at different directions along which one can travel and where one can end up. A list of alternatives is not a list of directions. An alternative is one state of affairs, a direction rather a path formed of them.

Some alternatives are close to those courses of development that we have already encountered and by which we are carried along, or along which we travel. Some alternatives, for their part, require large changes in order to become directions.

In addition, knowledge concerning the future almost always has some component relating to values and to justice. The space of possibilities cannot really be understood if one does not see the relation of the different "neutral" states of affairs to values — the capacity for normative thinking is essential when futures are studied. The identification and articulation of these values and of matters relating to justice is an essential part of futures studies. Some futures are desirable but difficult to attain, others are even probable but undesirable. Moreover, the desirability of a future is an inseparable part of how much we do towards it, of whether that future is to be attained. Our values thus affect the future causally. Studying the space of possibilities means articulating these alternatives, conditions and relations — the study not only of states of affairs but also of values. (See e.g. Virmajoki, 2026b.)

The space of possibilities is multidisciplinary, and knowledge concerning it combines different matters with one another — a representation of the future is indeed a kind of bundle of matters to be represented. For example, the future of education cannot be understood merely in terms of the development and use of pedagogy. Related to it are technology, working life, the economy, the everyday life of families, regional inequality, language, culture, politics and conceptions of the good life. All these must be included when we state that the claim "that future of education is possible" is knowledge about the future. Likewise the future of climate change is not merely a natural-scientific question, even though natural-scientific knowledge is necessary in it. Related to it are also energy, justice, consumption, international relations, technology, and political will. Knowledge of the future of climate change and of its effect is knowledge that gathers together the possible courses of development of these matters so as to describe one possible future; one point in the space of possibilities to which climate change is linked.

One must, however, be careful. The task of futures studies is not, and cannot be, to replace other disciplines. For example economics, climate science, sociology, education science and technology studies each produce important knowledge already as such but also for the sake of the future. The task of futures studies is rather to ask how the knowledge produced by different fields relates to possible futures and how different courses of development are linked to one another. Futures studies does gathering and articulating work when it produces knowledge of the possibilities of the future (see e.g. Bell, 2009; Virmajoki, 2026b).

Let it be noted that this is one reason why futures studies is often methodologically manifold. If the object of study consists, among other things, of economic, ecological, technological, cultural and political dimensions, it cannot be understood by means of one measure or one method. A complex object requires several viewpoints and several ways of producing knowledge, that is, methods.

Not everything possible stands on the same line

When one speaks of possibilities, it is important to distinguish different kinds of possibility.

A few examples: Not everything possible is probable. Not everything probable is desirable. Not everything desirable is easily realisable. Not everything possible is credible — something can be far-fetched in the light of our present knowledge, but we cannot rule it out. For example, in a situation of very great risk, [to take into account] a possibility that cannot be ruled out, even though there are no direct grounds for its realisation, is wise.

Somewhat more concretely, we can note, for example, that some future may be technically possible but politically difficult. Another may be economically attractive but ecologically unsustainable. A third may be ethically desirable but, in the light of present institutions, improbable. A fourth may be far-fetched on account of political energy dependencies, but still to be kept in mind, because some actor may cease to care about those dependencies.

The task of futures studies is to make such differences visible. It gives knowledge not only about the possibilities of the future as such, but also about the different kinds of grasping of the future.

This is why concepts matter. Claims concerning the future are not all alike. Roughly one can say, for example:

• A forecast strives to assess probable development.
• A scenario describes a possible future and a credible course of development.
• A vision describes a to-be-aimed-at future.
• A threat image raises to attention a to-be-avoided development.
• A roadmap describes the steps by means of which different futures can be reached.

If such kinds of knowing of the future are mixed together, "futures talk" easily becomes unclear. A scenario can be mistakenly assessed as though it were a forecast, even though its task would be to open up alternatives. A vision can be presented as a neutral description of what is to come, even though it contains values and aims. A threat image can be used as though it were a certain forecast, even though its task would be to help in preparing and not to compel action. Worth mentioning is, for example, that especially in the scenario literature it has been emphasised that the value of a scenario relates to credibility, coherence and the opening up of alternatives, not merely to the forecasting of realisation (see e.g. Ramírez & Selin, 2014; Schoemaker, 1995). And a scenario is knowledge. It is knowledge of a possible future. But only when it is justified credibly.

This can be illustrated with climate policy. An emissions scenario is not necessarily a forecast of what will certainly happen. It can describe what would happen if certain decisions were made now — even hard decisions. A vision of a sustainable society, in turn, is not a neutral forecast, but a target state — but it can be mere wishful thinking, if not even hard decisions can attain it. An assessment of climate risks, for its part, can describe to-be-avoided futures — and here the ghastliness of the risks must be taken into account, if certain decisions are not made. All these are needed, but they should not be assessed in the same way.

The same applies to an everyday example. If a young person is pondering a field of study, they can ask in which fields there will probably be work. They can also ask what kind of work they hold to be meaningful. That is a question relating to values. In addition, they can consider what kinds of new fields technology or societal change may bring into being. That is a question concerning possibilities. The decision, whether it turns out to be good or bad, does not arise from one answer, but from understanding the differences between the questions and their background assumptions.

Methods contain different assumptions about knowledge

In futures studies many methods are used, and some call it, some even with concern, a "method science." The manifoldness of methods can indeed at first appear confusing. Why are trend analysis, the Delphi method, scenarios, backcasting examination, examination of the operating environment, morphological analysis and critical layered analysis (CLA) needed? Why has an unambiguous method not been developed by which the relevant matters concerning the future could be distinguished and analysed?

The answer, of course, is that different methods answer different questions. A certain pluralism, the acceptance that there are different fields of study and within them different approaches, is already a generally recognised phenomenon (see e.g. Kellert et al. 2006). It is likewise in futures studies. As was set out above, not only one kind of thing is asked about the future. Even more important is that every method contains assumptions about why the present moment tells us something about the future, i.e. why it produces knowledge about the future. A method is never merely a technical tool. It is also a way of thinking the relation of knowledge, (un)certainty and action. For this reason the choice of method is always also an epistemic choice: when a method is chosen, it is at the same time decided what kind of knowledge concerning the future is held to be important in the situation (see e.g. Virmajoki, 2026a).

Different methods answer, for example, questions such as these:

• Trend analysis asks where present development would lead if it continued.
• The Delphi method asks for the assessments of experts as to what they hold to be possible, probable or important.
• Scenario work asks what the possible and credible courses of development are.
• Backcasting examination asks how a to-be-aimed-at future can be reached by returning (in the reasoning) from the future to the present moment.
• Examination of the operating environment asks what the emerging phenomena, weak signals and first signs of change are.
• Morphological analysis asks what possible combinations of factors there are.
• Critical layered analysis asks how worldviews, narratives and metaphors delimit what we think about the future.

Let us examine some methods more closely, so that we get some conception of the epistemic background assumptions. Behind the Delphi method lies the assumption that experts have knowledge, experience and a capacity for judgement and analysis that cannot wholly be reduced to statistics or models. The method can help to gather scattered knowledge, but at the same time one must ask who are chosen as experts and what kinds of viewpoints are thereby left out of the examination. Who is an epistemic authority? In the literature concerning the Delphi method it is precisely the gathering of expert assessments, the multiple rounds and the use of feedback between rounds that are the central features of the method (see e.g. Linstone & Turoff, 1975; Linturi & Kuusi, 2022). These assumptions rest on the fact that knowledge of earlier answers, and its gathering and iteration, are relevant when conceptions, knowledge, about the future are produced.

Behind scenario work, in turn, lies the assumption that, instead of one forecast, it is worth studying several alternatives. Good scenario work helps to see that the future is not merely a straightforward continuation of the present moment — we know that different events break the straightforward paths. In the literature concerning scenario methods it is precisely the construction of alternative courses of development, the articulation of uncertainties (which is an epistemic category) and the broadening of strategic thinking (i.e. the attaining of an end by means of knowledge) that are emphasised (see e.g. Bishop et al., 2007; Schoemaker, 1995).

Behind examination of the operating environment, again, lies the thought that significant changes often appear first as small and unclear signs: as weak signals, new discussions, deviations, experiments or slowly strengthening phenomena. Such crumbs of knowledge, as they are customarily called in Finnish, are regarded as significant. Examination of the operating environment does not tell which future will be realised, but helps to perceive the ingredients of change in the present moment — it guides us to produce knowledge of smaller but possibly decisive totalities (see e.g. Cuhls, 2020; Dufva, 2022; Hiltunen, 2008).

Critical layered analysis (CLA: Causal Layered Analysis) goes in yet another kind of direction. Behind critical layered analysis lies the thought that the deeper assumptions of images of the future can be studied in layers, from surface phenomena to worldviews and metaphors (see e.g. Inayatullah, 1998). It asks what kinds of conceptions of systems, what kinds of worldviews, narratives and what kinds of metaphors delimit what we hold to be possible — it divides human thinking and knowledge into layers that affect one another and thereby the human capacity to see futures. If I think I know that the human being is always selfish (worldview), I also think I know that the environment cannot be saved, because incentives create a system in which nature is left secondary. The "knowledge" of one level leads to the "knowledge" of another level.

The manifoldness of methods is thus not confusion or a fragmentation that could be corrected. It is a consequence of the fact that there are many kinds of questions concerning the future. Good futures studies does not choose a method by chance. It asks, epistemically speaking, what a certain method assumes knowledge to be, who knows, and where knowledge is located.

Models and scenarios are not copies of reality

At this point the reader may ask: if scenarios are not forecasts and models are simplifications, why should they be held to be "knowledge"?

The question has risen, at a general level, to the centre of the philosophy of science, not merely in the case of knowledge concerning the future. In science, models that are not perfect copies of reality are used continually. Models delimit, simplify and emphasise something essential. Their purpose is not always to describe everything, but to help in explaining and understanding something. In the philosophy of science concerning the epistemic value of models and idealisations it has been argued how a model can be useful precisely because it makes some structure, relation or feature understandable (see e.g. Elgin, 2017). It is not, for example, essential which is the highest mountain in order to understand why the Earth orbits the Sun along a certain path — a question that was at the centre of the birth of modern science.

Let us take an everyday example. A map does not contain knowledge of everything that is in the terrain. A good road map leaves out an enormous amount of knowledge: trees, the details of buildings, sounds, smells and so on. Still it is an extremely good aid when moving about. In the same way a scientific model can leave out many things, if it helps to understand a certain dependence or structure. A model is bad when it leaves out something essential in relation to the purpose for which it is used. If the highest mountain were to affect the Earth's orbit essentially, leaving it out of account would be a mistake.

Scenarios function partly in the same way. They are not exact pictures of possible futures, but articulated descriptions of possible courses of development. A good scenario does not try to say everything about its object. It brings out something essential: dependencies, uncertainties, points of choice, risks or possibilities. In addition it must be noted, as above, that the quality of a scenario is not assessed only on the basis of realisation, but also on the basis of credibility, coherence and strategic usefulness (see e.g. Ramírez & Selin, 2014; Schoemaker, 1995). When the dependencies, uncertainties, points of choice, risks and possibilities are not credible, coherent or strategically useful, the scenario is bad.

Values are not a disturbance, but part of the research

As we have seen, in knowledge concerning the future values come in unavoidably. The centrality of values is not a weakness of futures studies, but rather a part of its object of study.

If we ask what a good future of education would be, we do not ask about the state of affairs as such. We also ask what good learning, cultivation, equality, efficiency or justice would be. If we ask what a sustainable economy would be, we must ask what is meant by sustainability and from whose viewpoint. We must know what values are to be taken into account when the future is examined. Knowledge must be produced of values and of possible values.

Generally speaking, the role of values in science is more broadly important as well. Values can affect, for example:

• what is held to be worth studying
• what kinds of risks are held to be serious
• what kind of evidence is held to be sufficient
• how research results are interpreted
• to what research knowledge is put

This does not mean that scientific knowledge could always be accepted or rejected purely on the basis of values. Rather, good research makes its background assumptions and its values visible and does not hide them behind an apparent neutrality and thereby make the same mistake, only in hiding (that is, an acceptance/rejection based on values). (See e.g. Douglas, 2009; Longino, 1990.)

In futures studies this is emphasised, because the research often relates to decisions — often even more than in the case of other fields. When a vision is drawn up, it is asked what is aimed at. When a backcasting examination is made, some desired future is first chosen. When risks are assessed, it must be asked what is held to be a serious harm. When scenarios are constructed, it is decided which factors, uncertainties and actors are raised to attention. The significance of values in futures studies has indeed been held to be an essential part of the nature of the field, not merely an addition external to the research or a matter that must be assessed from outside (see e.g. Bell, 2009).

The involvement of values does not, then, make futures studies unscientific. It would rather be unscientific to present value choices as pure facts. Good futures studies tells openly when it anticipates, when it assesses, when it aspires and when it warns.

This is especially important when knowledge concerning the future affects other people. If an image of the future concerns education, it concerns pupils, teachers, families and society's conception of cultivation. If some knowledge concerns the climate and its possible future, it also concerns coming generations. If knowledge concerns technology, the possible effects on work, privacy, security and power must be known. Claims about the future are not merely thoughts. They are knowledge in which values are present both as object and as part of the structure of the research itself (as in other fields of study too).

Knowledge concerning the future is special also because it can affect the present moment, which in turn affects the future. The forms of futures knowledge — such as forecasts, scenarios, visions and threat images — do not remain merely on paper. They can steer decisions, investments, policy and people's expectations. If some technology is presented as inevitable, companies and decision-makers may begin to act accordingly. If some risk is described as serious, it can lead to regulation and preparedness. This can be called reflexivity.

Claims and descriptions concerning the future are not merely external representations of possible futures. They also link up as part of the action of the present moment. People make decisions on the basis of what they hold to be possible, probable or desirable. These decisions in turn change what kinds of futures are possible. In the discussion concerning anticipation and futures literacy it has indeed been emphasised precisely that conceptions concerning the future are part of action, or will come to be formed as part of action, in the present moment (see e.g. Miller, 2018).

For this reason futures studies must be especially responsible also with regard to knowledge. A poorly justified threat image can cause fear or steer towards bad decisions. Too certain a technology forecast can narrow thinking. On the other hand, well-done futures studies can open up alternatives, help in preparing, make values visible and increase the capacity to act amid uncertainty. It is not a question of whether the future affects itself, figuratively speaking, through knowledge and values, but of whether the knowledge is justified and the values acceptable.

When is knowledge concerning the future good?

We have now dealt with what kind of knowledge futures studies produces — its conceptions of knowledge — and with knowledge as part of the research. It remains only to ask what this discussion means — different conceptions of knowledge do not yet tell what knowledge is of good quality / acceptable or of poor quality / to be rejected.

Good knowledge concerning the future is at least:

• Clear about its nature. The recipient must know whether it is a matter of a forecast, a scenario, a vision, a warning, or whatever each thing generally is.
• Justified. Behind the claims there can be material, expert assessments, theory, modelling, participants' experiences or value analysis. The justifications must be in view.
• Coherent. The parts of the description of the future must fit together. Coherence does not make a scenario good as such, but it distinguishes a scenario to be taken seriously from a random and confused narrative.
• Linked to the present moment. A credible image of the future must have some traceable connection to present phenomena, trends, decisions, structures or possible changes at hand.
• Opening up of uncertainty. Good futures studies does not pretend to certainty where there is no certainty. Nor does it use uncertainty as a pretext for defending any claim whatever.
• Responsible. Knowledge concerning the future can affect people's possibilities, rights, security and decision-making. Its production and communication must therefore be ethically upright.

These criteria do not form a list by means of which it can be mechanically checked whether some research on the future was of good or of poor quality. They are rather ways of asking whether knowledge concerning the future has been produced carefully and in a principled way. Good futures studies does not claim to know everything. It tells what it knows, what it assumes; what it does not know, and when even ignorance can have significance.

The future is not known, but it can be understood

The core of the concept of the knowledge produced by futures studies can ultimately be expressed simply:

• The future is not known in advance, but the possibilities opening up from the present moment can be studied systematically by different methods for different purposes.

Futures studies does not know what will altogether certainly happen — do we not often fail to understand even the present moment in our own lives? The knowledge of futures studies is knowledge, because it helps to see what could happen, why, under what conditions and with what kinds of consequences. Futures studies produces understanding: the capacity to act on the strength of what is known about relations of dependence, states of affairs, background circumstances and actors. It helps to understand how movement towards a certain future is brought about and how it is influenced.

For this reason futures studies needs both observations and imagination, both models and reflection on values, both expert knowledge and critical discussion. It is not mere forecasting, but neither is it free imagining with one's head in the clouds. Its task is to help people and communities to discern the space of possibilities in which the decisions of the present moment are made.

Precisely in this lies the epistemic significance of futures studies. It does not give certain knowledge about the future. It gives a better way of asking, thinking and acting when certain knowledge is not to be had.

References

Bell, W. (2009). Foundations of futures studies: Human science for a new era. Transaction Publishers. (See also vol. 2 of the two books.)

Bishop, P., Hines, A., & Collins, T. (2007). The current state of scenario development: An overview of techniques. Foresight, 9(1), 5–25.

Cuhls, K. (2020). Horizon scanning in foresight: Why horizon scanning is only a part of the game. Futures & Foresight Science, 2(1), Article e23.

de Regt, H. W. (2017). Understanding scientific understanding. Oxford University Press.

Douglas, H. (2009). Science, policy, and the value-free ideal. University of Pittsburgh Press.

Dufva, M. (2022). Toimintaympäristön analyysi: PESTE ja sen variaatiot. In H.-K. Aalto, K. Heikkilä, P. Keski-Pukkila, M. Mäki & M. Pöllänen (eds.), Tulevaisuudentutkimus tutuksi – perusteita ja menetelmiä. Tulevaisuudentutkimuksen Verkostoakatemia.

Elgin, C. Z. (2017). True enough. MIT Press.

Hiltunen, E. (2008). The future sign and its three dimensions. Futures, 40(3), 247–260.

Inayatullah, S. (1998). Causal layered analysis: Poststructuralism as method. Futures, 30(8), 815–829.

Kellert, S. H., Longino, H. E., & Waters, C. K. (Eds.). (2006). Scientific pluralism (Minnesota Studies in the Philosophy of Science, Vol. 19). University of Minnesota Press.

Linturi, H., & Kuusi, O. (2022). Tulevaisuuksia ennakoiva Delfoi-menetelmä. In H.-K. Aalto, K. Heikkilä, P. Keski-Pukkila, M. Mäki & M. Pöllänen (eds.), Tulevaisuudentutkimus tutuksi – perusteita ja menetelmiä. Tulevaisuudentutkimuksen Verkostoakatemia.

Linstone, H. A., & Turoff, M. (eds.). (1975). The Delphi method: Techniques and applications. Addison-Wesley.

Longino, H. E. (1990). Science as social knowledge: Values and objectivity in scientific inquiry. Princeton University Press.

Miller, R. (ed.). (2018). Transforming the future: Anticipation in the 21st century. Routledge/UNESCO.

Ramírez, R., & Selin, C. (2014). Plausibility and probability in scenario planning. Foresight, 16(1), 54–74.

Schoemaker, P. J. H. (1995). Scenario planning: A tool for strategic thinking. Sloan Management Review, 36(2), 25–40.

Virmajoki, V. (2026a). The shining: Illuminating philosophy and futures studies. Futures, 179, Article 103812.

Virmajoki, V. (2026b). The greatest mystery of futures studies: The possibility space as a research object. Futures & Foresight Science, 8(2).

Woodward, J. (2003). Making things happen: A theory of causal explanation. Oxford University Press.

Philosophy has a bad reputation in many places, and sometimes for a good reason. When people hear "philosophy of the future," they expect speculation: what is time, does the future exist, can we ever know anything. These are old questions and they will stay unresolved. But philosophy, as it is practiced today in philosophy of science, is not that. It does not speculate. It looks at what researchers actually do and asks concrete questions about it.

Here I must be honest: the futures field itself carries part of the blame. When futures studies does "philosophy," it too often means exactly the speculation described above, or debates between grand frameworks – whether one theory "subsumes" another, whether the practitioner has reached the right developmental stage of consciousness to use a method properly. One can read whole special issues of this. The debates generate heat but no light. They do not sharpen any concept and they do not help anyone to do better research. And at the same time, the field keeps inventing the wheel again: it struggles with problems – how to justify claims about things that cannot be observed, how methods relate to evidence, how values enter into research – that philosophy of science has worked on for decades with quite sophisticated results. Historians, astrophysicists and evolutionary biologists also study what they cannot directly observe. The tools exist. They wait for use.

So what philosophy actually asks, then? It asks how a field of research works in practice: what are the goals, the methods, the objects of study, the theories. It asks what tensions there are – for example, futures studies produces multiple scenarios because the future is uncertain, but policy-makers want one reliable forecast. That is a real tension in the working of the field, and it can be analyzed. It asks how methods are justified. Not in some absolute sense, against the universe, but relative to a goal: does the method deliver what it promises for the purpose it is used? A method that does not predict anything is not a failed method, if prediction was never its job. And it asks what assumptions are built inside the methods, because there always are some, whether we like it or not.

Take Causal Layered Analysis. Philosophical analysis of CLA does not ask whether poststructuralism is "true" or whether the layers "really exist" in some deep metaphysical sense. It asks: when a researcher identifies a worldview or a myth in the material, what exactly they are picking out? How the interview answers get translated into claims about deeper cultural patterns? These questions are not attacks on the method. They make the method more understandable and, possibly, better.

This kind of philosophy is modest and, to be honest, quite dull. But it is no bullshit. Every question has a connection to actual research practice. The field does not need more visionaries of ultimate reality; it needs this dull, careful work. It is better to know that something cannot be done than to not know and make futile attempts.