Fine and Rosen: A Note on Metaphysics, Logic and Epistemology in Philosophy of Education (4)

Up to now I have been asking, first, how to distinguish necessity from essence, then how ordinary modality relates to more formal or special modalities, and then how those distinctions might help educationalists think more carefully about capacities, constraints, sequencing, space, timing, and institutional design. Gideon Rosen forces a sharper question. It is not just, what is essential to a thing, or what can or must happen under certain conditions, but: what kinds of essences are even legitimate candidates for there being? Which alleged natures are genuine, and which are spurious? Which definitions carve out something real, and which merely combine words without corresponding to any proper item, property, or educational form at all?

Educational discourse is crowded with candidate essences. We speak as if there is a thing called critical thinking, creativity, mathematical ability, classroom order, school readiness, disciplinary knowledge, inclusion, leadership, resilience, engagement, curriculum coherence, educational progress, even the learner as such. But what gives us the right to suppose that, for each such candidate, there is a real property, practice, or institutional form whose nature can be articulated? Why is “being a well-ordered classroom” a more plausible educational kind than, say, “being a lesson in which Jane suddenly becomes curious at 10:17,” or “being a curriculum whose hidden essence is that pupil X will one day be a banker?” 

Rosen’s problem is exactly about this, though stated in general metaphysical terms. For which propositions is there something whose essence they are? What can legitimately go into the essence of a property or an individual? Educational theory very often moves too quickly from a useful description to an alleged ontology. A pattern is observed, a discourse gains traction, a policy category becomes administratively central, and before long it is treated as if it names a genuine property with its own determinate nature. The Rosen question gives us a way of slowing that movement down. It suggests constraints on essentialist existence. Not every coherent sentence, not every practically useful classification, and not every policy label should be taken to correspond to a genuine essence. There may need to be principled restrictions.

The issue for Fine is not essential existence, not which things exist essentially, but what he calls the problem of essentialist existence, which candidate essences exist at all? Given some proposition p, when is there a property F such that it lies in the nature of F that p? In the notation Fine uses, when is there an F such that □F p. This is a question about the legitimacy of essence-attributions themselves.

The educational parallel is immediate. Suppose one proposes that there is a property of being a successful learner such that it lies in its nature to persist through challenge, respond to feedback, self-regulate, transfer skills, and maintain motivation. Another proposes that there is a property of being school-ready such that it lies in its nature to sit still, wait one’s turn, decode print, and follow institutional rhythms. A third suggests that there is a property of being authentically creative such that it lies in its nature to resist all curricular scripting. The Rosen Fine question is not yet whether these are true descriptions. It is whether there is even a legitimate educational property here whose essence could be so specified. 

Just as one cannot suppose that every condition determines a set, one should not suppose that every proposition determines an essence. In set theory, Russell’s paradox showed that naive comprehension fails. Not every intelligible condition gives rise to a legitimate set. Likewise, in essentialist theory, not every intelligible proposition should be expected to give rise to some property or individual whose nature it is. This is useful for educationalists because education is full of what we might call naive essentialism. It often assumes that because we can formulate a pedagogic category, there must be a corresponding educational essence. Fine suggests caution. There may need to be constraints analogous to those that disciplined set theory.

One can now see the relevance of explanatory challenge, which Fine stresses. Rosen’s toy example is that there could not be a property Z such that it lies in the nature of Z that Jane is a banker. Fine agrees that presumably there could not be such a property, but insists that if not, there should be an explanation. It should not be a mere accident that some proposed essences are illegitimate. Why is “being a numerate child” a more legitimate candidate property than “being a child who will eventually pass GCSE maths in 2032?” Why is “being a history curriculum organised around disciplinary enquiry” a more serious educational kind than “being the curriculum whose essence is that classroom 4B smells faintly of glue and contains two distracted pupils?” Intuitively, the latter sorts of properties seem spurious or at least non-educationally real. But what explains the difference? Fine’s framework encourages us to seek principled answers rather than relying on taste or disciplinary habit.

The first major route Fine explores is explicit definition. This is the cleanest case. A property F is explicitly defined when we can give necessary and sufficient conditions for being F in a non-circular way. In the simplest formal case, this takes the form that for all x, Fx if and only if φ(x), where φ does not itself contain F or anything dependent on F. Here the property’s nature is explicitly given in terms of other properties. Fine treats this as a canonical form of definition.

Educationally, many of our strongest categories aspire to something like explicit definability, even if they rarely achieve it cleanly. Consider “being a prime number” in mathematics education, or “being a finite verb” in grammar, or perhaps “being formative assessment” in educational theory if the concept were used with unusual discipline. We try to give criteria that are meant to be both necessary and sufficient. If a property is explicitly definable in this way, Fine argues, then we have not only a reason to think such a property exists, but also a reason to think that its explicit definition exhausts its essence. There is, so to speak, nothing more to being that property than what the definition gives.

This idea of exhaustive essence is extremely fertile for educational modelling. There are some educational categories whose natures may be close to exhaustive in this way, especially technical or local ones. A timetable clash, for example, may be explicitly definable in terms of two scheduled commitments occupying the same time slot for the same agent. Once that is stated, little more remains to the nature of the thing. Likewise, a prerequisite relation in a tightly formal domain may in some cases be explicitly characterisable. If these categories are explicitly definable, then their educational use can be sharpened. We can identify sufficient conditions for their presence and clearer limits on what further properties can legitimately be inferred.

But Fine is equally alert to the limits of this route. Many properties are not explicitly definable. Indeed, if every property had to be explicitly defined in terms of another, one would face regress. So explicit definition cannot be the whole story. And I'd say that some of the central categories we care about, understanding, knowledge, interpretation, authority, inclusion, perhaps even teaching itself, do not look explicitly definable in neat necessary-and-sufficient terms. Yet it would be perverse to conclude that they therefore lack genuine natures altogether. So we need a more flexible approach.

Fine makes an important distinction between constitutive and consequentialist notions of essence. Under a constitutive conception, the essence of a thing is given by its basic defining truths. Under a consequentialist conception, the essence includes whatever follows in the right way from those defining truths. Fine works, at least for present purposes, with a consequentialist notion. This means that if it lies in the nature of F that all and only x such that φ(x) are F, then it may also lie in the nature of F that all F are φ, that some weaker consequences hold, and that truths mediated through the essences of dependent properties can also belong to F’s essence.

Educational understanding and educational ontology are almost never exhausted by bare defining biconditionals. Once a practice or property has a certain nature, many consequences flow from that nature. If the nature of a seminar includes reciprocal attention, responsiveness to reasons, and sustained thematic continuity, then many further truths may follow about turn-taking, silence, interruption, preparation, and authority. A consequentialist approach allows us to say that such downstream truths belong to the essence in an extended way. I think educational practices are often like this. Their natures ramify.

Fine then develops the notion that an explicit definition can exhaust the nature of a property. He formulates a necessary condition according to which, if F is explicitly defined by φ, then anything essential to F must follow, in the appropriate mediated way, from that definition together with the essences of the properties on which φ depends. In other words, the definition exhausts the essence. 

For philosophy of education, this suggests a valuable discipline. When educationalists propose that some category has a determinate nature, they should ask whether the proposed basis really exhausts the kind. If one defines formative assessment simply as assessment used to adapt teaching in real time, does that exhaust its essence? If not, what more belongs to it, and from where does that additional content come? If one defines classroom order simply as the absence of disruption, is that truly exhaustive, or does order also include positive norms of attention, trust, and shared orientation? Fine’s framework pushes us to distinguish between categories whose natures are exhausted by a formal role and categories whose apparent definitions do not get us far enough.

He also notes that explicit definition may yield what he calls essentialist islands. Explicitly definable properties can be relatively self-contained. Their essences do not ramify arbitrarily into the essences of everything else. This is a useful thought for educational design. Some parts of educational systems can and perhaps should be treated as relatively formal islands, punctuality rules, attendance coding, timetable conflict, room capacity, examination scheduling. These need not be denied complexity altogether, but their local ontologies can be reasonably tight. Other parts of education, understanding, authority, belonging, aspiration, discipline, cannot be treated in this way without distortion. They are not islands. They leak out.

The next step is implicit definition, and here the discussion becomes even more interesting for our purposes. Some properties seem to have natures even though they cannot be captured by any explicit non-circular necessary-and-sufficient condition. Fine mentions knowledge as the kind of case Rosen has in mind. It may lie in the nature of knowledge that if S knows that p, then p is true, even if knowledge is not explicitly definable. Many of our central educational concepts are like this. We may have some confidence that understanding, attention, curiosity, disciplinary judgement, classroom authority, educational trust, or intellectual seriousness have genuine natures, but not think they can be neatly defined. They are not therefore formless. The question is how to model such cases without collapsing into vagueness or arbitrary stipulation.

Fine’s proposal suggests a conservativity condition. If a supervening or dependent property has a certain essence involving only its subvening or base properties, then that truth should already be an essence of those subvening properties. Introducing the higher-level property should not create new truths about the base. In his phrase, a supervening property should not, by its very nature, tell us anything more than what was already implicit in the nature of the subvening properties. This is an extremely promising principle for educational ontology. Suppose someone introduces the property of being a well-managed classroom, and claims that it lies in its nature to have orderly transitions, low-level attentiveness, stable norms, and reciprocal recognition of authority. Fine’s conservativity thought would suggest that if this property is legitimate, then these truths must already be implicit in the nature of the lower-level practices, norms, relations, and spatial-temporal arrangements on which it depends. The supervening educational property should not conjure substantive new structure from nowhere. It should gather, organise, or witness a structure already there.

This gives us a principled way to resist some inflated educational categories. Consider the property of being “future-ready,” a phrase often used in policy. If this property is treated as if it by nature includes adaptability, creativity, collaboration, resilience, digital fluency, and entrepreneurial spirit, the conservativity demand asks whether these are genuinely implicit in the lower-level educational realities or whether the label is illicitly creating an aura of coherence where none exists. Likewise with “engagement.” Does positing engagement as a property explain anything, or merely repackage a heterogeneous bundle of more basic relations, behaviours, and conditions? Fine gives us a test. A legitimate higher-order educational property should be conservative with respect to its base.

He then pairs this with a kind of witnessing principle. Very roughly, if it lies in the nature of some base properties that there is a property playing a certain role, then there is such a property, one dependent on that base, whose nature it is to play that role. This is a strong claim and Fine acknowledges as much. But restricted to suitable cases, it is very suggestive. In educational terms, this means that if the structure of certain base practices and relations really calls for a higher-order educational category, then positing such a category is legitimate. For example, if the nature of certain discursive, disciplinary, and normative practices makes room for a property we might call historical thinking, then there may be a genuine educational property there, dependent on those more basic structures, whose nature is to play that role. This gives a principled route for educational concept formation. Categories need neither be arbitrary stipulations nor mysterious discoveries. Some may be witnessed by the organised structure of more basic educational realities.

Such witnessing should not be promiscuous. Fine himself worries about strong forms of the principle. That is important educationally too. One cannot infer from every pedagogic pattern that there must be a corresponding educational property. Otherwise one generates artificial hybrids that do no real explanatory work. So the question becomes, in which cases do our educational bases genuinely witness a higher-order property, and in which cases are we merely projecting administrative or rhetorical categories onto practice?

Rosen presents the guardian angel problem. If it can lie in the nature of singleton Socrates to be a set that exists whenever Socrates exists, why can it not lie in the nature of Socrates’ guardian angel to be an angel who accompanies him whenever he exists? Why does one essence seem legitimate and the other not? Fine’s answer turns to generative ontology. Some entities are generated through certain generative relations, sethood and membership being the paradigm. There is a cumulative ontology of generated entities and an associated ideology of generative relations. Sets are generated from their members; certain entities are introduced at stages of a hierarchy according to the definitional and dependency relations that govern them. The key thought is that some properties and relations are generative in character. Their nature consists in their role in generating entities. Others, like angelhood and guardianship, are not generative in that way.

This suggests that some educational entities and structures are legitimate because they are generated by, and exhausted by, the constitutive relations of the practice, while others are not. A timetable slot, a group roster, a mark distribution, an assessment profile, a prerequisite chain, even perhaps a lesson sequence, may be more like generated entities. Their existence is not a further substantial fact over and above the organised relations that constitute them. They are educationally insubstantial in Fine’s sense, not unreal, but not adding a further worldly layer beyond their generative structure. By contrast, positing some educational individual or property may involve a further substantial fact that is not generated by the relevant practice-constituting relations. Think of claims like “every learner has a unique inner academic destiny,” or “each child’s true curriculum awaits discovery,” or “every classroom has an invisible pedagogic spirit that follows it across institutional settings.” These may not merely be false. They may fail because the relations invoked are not of the right generative character. They do not define educational entities in the way membership defines a singleton or curricular sequencing defines an administrative placement. Fine’s framework gives us a way to say why some educational positings seem ontologically disciplined and others extravagant.

This also clarifies a danger in educational romanticism as well as in bureaucracy. Bureaucracy tends to generate pseudo-entities by administrative fiat, treating score profiles, progress bands, and intervention labels as if they named robust educational realities. Romantic discourse can do something analogous from the other side, positing deep educational essences for inspiration, voice, or becoming without asking whether the relevant relations actually generate such items. Fine’s generative ontology offers a discipline for both. It asks whether the candidate item is genuinely generated by the constitutive relations at hand, and whether positing it adds anything substantial beyond those relations.

His more general conservativity principle for individuals then adds a further layer. If it lies in the nature of an individual to stand in certain relations to other items on which it depends, then those truths must already be implicit in the natures of those other items. Again, the individual should not smuggle in new reality. This suggests that if one says, for example, that a school culture, or a classroom ethos, or a curricular pathway, is an individual-like educational entity, then the truths essential to it should be traceable back to the practices, norms, spaces, times, and relations on which it depends. This is a powerful check against reification.

At the same time, Fine’s treatment of generated entities helps us see why some educational wholes are not mere fictions. A classroom is not simply a list of pupils plus a room. It may be generated at a certain stage by a structured conjunction of timetable, teacher allocation, spatial arrangement, institutional norms, and reciprocal expectations. Its educational identity may therefore be real, though derivative. Likewise, a curriculum sequence may not be a mere aggregate of lessons. It may be a generated individual or structured entity whose nature is partly fixed by generative relations of order, dependency, and revisitation. This is a way of securing real educational ontology without either flattening everything into aggregates or inflating every useful category into an autonomous essence.

What, then, does this add to my larger discussion? It deepens and progresses it in several ways. First, it moves us from asking what is essential to educational things to asking which educational things and properties are legitimate bearers of essence in the first place. Educational theory often argues over the nature of categories before establishing whether the categories are ontologically disciplined. Second, it gives us a richer account of different routes by which educational kinds might be legitimate. Some may be explicitly definable and relatively exhaustive. Some may be implicitly definable and conservative over their bases. Some may be generated by constitutive relations and therefore be derivative but real. This is far more nuanced than the flat choice between realism and anti-realism that often governs educational debates.

Third, it provides a principled way to resist both inflation and reduction. One need not reduce every educational category to psychology, sociology, or administration. But neither may one posit robust educational essences at will. The notions of exhaustive essence, conservativity, witnessing, and generative character offer middle paths.

Fourth, it helps connect metaphysical modelling to educational practice with greater precision. When thinking about classroom management, curriculum, architecture, or assessment, we can now ask not only what modalities are in play, or what grounds what, but also what kinds of entities and properties the practice legitimately generates. Does a behaviour system generate a real educational order, or merely a coercive surface regularity? Does a curriculum map generate a genuine sequence, or only an administrative representation? Does an assessment rubric define a real property of performance, or merely impose a convenient classificatory overlay? These are Rosen Fine questions in an educational register.

Fifth, it gives a stronger answer to sceptics of metaphysics in education. The issue is not whether one likes abstract theory. The issue is that educational thought keeps generating and using categories, relations, and wholes. Without some account of which of these are legitimate, how they are defined, what they depend on, and whether they are substantial or derivative, educational discourse becomes ontologically irresponsible. Fine shows that these are not ornamental questions. They are conditions of serious conceptual work.

And finally, this discussion helps explain why educational reform so often fails conceptually before it fails practically. Reforms frequently introduce new entities, learner profiles, competencies, pathways, readiness levels, cultures, ecosystems, capabilities, without asking whether these are explicitly definable, implicitly witnessed, conservatively grounded, or genuinely generated by the practice. The result is often ontological clutter. Fine and Rosen suggest educationalists need to become more careful not only about evidence and value but about what sorts of things they are entitled to say there are. 

The earlier discussions were about starting to think educationally in terms of necessity, essence, modality, dependence, and ground. The Fine Rosen discussion adds a prior and regulating question, what educational essences and educational entities are even fit to be admitted into our ontology? Once that question is asked with sufficient seriousness, educational thought becomes at once more modest and more exact. It becomes less likely to reify passing abstractions, less likely to naturalise administrative categories, and more capable of identifying which educational forms are genuinely there to be understood.

You think at first it’s just a bit of philosophical tidying up, nothing dramatic, just making sure we’re not being sloppy. That’s the agreeable version. But then you actually try to use it, even in a very everyday way, and something slightly awkward happens. Some of the things you normally say, quite happily, start to feel a bit overconfident, a bit too neatly packaged, as if they’re pretending to be more solid than they are. Take the teacher. You say, this teacher is excellent. Or she has authority. Or she’s inspiring. It all sounds perfectly in order, almost reassuringly so. You can picture it, you’ve seen it, it’s familiar. But then you stay with it, just for a moment longer than usual. Authority, for instance. It feels like something the teacher has, a kind of presence, something you could almost point to. And when it’s there, it’s very convincing. The room settles, things click into place, there’s a sense that this is working. But then, slightly disconcertingly, you see it not working. Same teacher, broadly speaking, same way of speaking, nothing obviously wrong, and yet the authority has thinned out, or just isn’t there. It’s not a collapse, more a kind of absence. And that’s the moment where the original description starts to wobble a bit.

Because if it were really just something the teacher had, it shouldn’t do that. It shouldn’t come and go in quite that way. So you begin to think, well, perhaps it isn’t located there, not in that simple sense. Perhaps it’s something that happens, which sounds vague, but isn’t meant to be. It happens when certain things line up, the students’ responses, the expectations in the room, what’s been established before, all the small, slightly unglamorous elements that don’t sound like “authority” but seem to be doing the work. None of them is authority on its own, obviously. But without them, nothing. So it’s real, clearly, but it’s not self-sufficient. It doesn’t stand there on its own, however much we’d like it to.

Which makes the idea of training someone to “have authority” feel just a bit optimistic. Not entirely wrong, but as if it’s pitched at the wrong level. As if you’re polishing the surface and hoping the rest will follow.At the same time, it isn’t just behaviour either. You can have all the right gestures, the right tone, and still it doesn’t quite come off. So it’s neither an inner possession nor a visible technique. It’s something more awkward than that, something distributed, if you like, across the situation.

“Inspiring” is even more seductive, because it sounds like a deep personal quality. This teacher is inspiring. It feels like a compliment with substance. But again, you watch it shift. Another class, another moment, and it isn’t quite there, or not in the same way. And it doesn’t feel like a contradiction, more like you’ve been a bit too quick off the mark. Because what you’re calling inspiring seems to depend on quite a lot, the material, the timing, whether the students are receptive, whether attention gathers or just drifts. When it works, it really works, there’s a kind of lift to it. When it doesn’t, it’s oddly flat. So again, it’s real, but it isn’t a fixed feature sitting inside the person.

Then the class, which people talk about as if it were a thing in its own right. This class is difficult. This class is engaged. And you know exactly what’s meant, there’s a definite atmosphere, almost a personality. But if you try to pin it down, it becomes slightly elusive. You can have a room where everyone is behaving, doing what they should, and yet there’s no sense of anything happening collectively. It feels inert. No buzz. Then another room where something clicks, there’s a rhythm, people anticipating, responding, and it feels like a single thing, even though it obviously isn’t. So the class seems to be something that emerges, which sounds like a vague word but is actually quite specific. It emerges from the way the parts relate, over time, with the teacher, with each other, with the routines that have settled in. Not just the sum of individuals, but not independent either. When you say this class is engaged, you’re not just totalling up attention. You’re pointing to that pattern, that shared orientation. It depends on individual attention, of course, but it isn’t reducible to it. Change the underlying arrangement and the engagement disappears, even if some of the outward signs remain.

“Culture” is where things get a bit inflated. Schools like to talk about building a culture, as if it were something you could more or less install. It sounds impressive. But unless you connect it to actual practices, repeated interactions, what gets noticed, what gets quietly discouraged, it starts to feel a bit like a slogan. Substantial sounding, but slightly empty. Once you tie it back to those patterns, it becomes more convincing. It’s there, real as can be, but only because those things are there. Without them, the word doesn’t quite stick.

Then the student, and the tone shifts again. This student can do algebra. This student cannot yet write analytically. It all sounds quite matter of fact. But it’s really about what is possible, and that’s where it gets interesting. Because “can” always seems to come with conditions, even if you don’t spell them out. Given what the student knows, given the support, given the task, then yes. Change those, and perhaps not. So the ability isn’t a fixed property. It’s tied to a situation, even if we talk as though it isn’t. That stops it becoming a kind of essence, something the student simply has. But it also doesn’t reduce to a one-off description. It’s still making a claim about what would happen, just not in an absolute way. It also helps separate things that tend to get muddled. Being able to do something and being required to do it aren’t the same. A student who doesn’t complete homework might be perfectly capable, or not. Without separating those, you get a slightly distorted picture. Ability itself starts to look less robust the more you look at it. It shifts with context, with training, with support. It doesn’t behave like something fixed, even though we often talk as if it does.

Assessment shows this quite starkly. A rubric says secure understanding, and it sounds like something the student now possesses. But what you actually have is a performance, in a particular moment, under particular conditions. To turn that into a stable feature of the student, you’re making an extra move, assuming it will carry across contexts. Sometimes it will, sometimes it won’t, but the assumption is doing a lot of work.

The same slightly overconfident tone appears in larger claims. A curriculum builds knowledge progressively. Sometimes that’s genuinely true, there’s a real dependency there. Sometimes it’s just an ordering that’s been imposed. The language makes it sound more necessary than it might be. Or take a space. An open-plan classroom promotes collaboration. It probably does, in certain ways. But it also makes other things harder. So the space is shaping what can and can’t happen, whether or not that’s acknowledged.

By this point nothing has disappeared. Teachers, classes, students, they’re all still there, still real, just as before. But they don’t feel quite as self-contained. Each one seems to depend on a whole set of conditions, slightly fragile, slightly contingent. And the effect is not dramatic, not a grand revelation, more a quiet adjustment. The confident, slightly grand way of talking about these things starts to feel just a bit too neat, a bit too pleased with itself. You can still say it, of course. It’s not forbidden. It just doesn’t quite ring true in the same way once you’ve seen what’s holding it up.

And that's when you start to think.