Modality and Variation

David Corfield suggests that modality has to do with ranges of variation. This seems extremely helpful. He connects Brandom’s notion of ranges of counterfactual robustness with mathematical analyses of variation. Corfield approvingly cites Brandom’s argument that in order to successfully apply empirical concepts at all, we must already be able to apply modal concepts like possibility and necessity. This always seemed right to me, but the talk about possible worlds made me worry about what sounded like impossibly strong quantification over infinities of infinities. Corfield also points out that Saul Kripke originally cautioned against uncareful extension of his possible-worlds talk.

It now seems to me that Brandom’s counterfactual robustness and Corfield’s mathematically analyzable variation together can be taken as an explanation for modal notions of necessity that previously seemed to be simply posited, or pulled out of the blue. Modality suddenly looks like a direct consequence of the structure of ranges of variation. Previously, I associated both structure and Brandom’s modally robust counterfactuals with Aristotelian potentiality, so this fits well.

Corfield also relates this to work done by the important 20th century neo-Kantian, Ernst Cassirer, on invariants behind the various systems of Euclidean and non-Euclidean geometry. He points out that Cassirer thought similar concerns of variation and invariance implicitly arise in ordinary visual perception, and connects this with Brandom’s thesis that modality is already there in our everyday application of empirical concepts.

The British Empiricist David Hume famously criticized common-sense assumptions about causality and necessity, preferring to substitute talk about our psychological tendencies to associate things that we have experienced together. Hume pointed out that from particular facts, no knowledge of causality or necessity can ever be derived. This is true; no knowledge of necessity could arise from acquaintance with particular facts. But if necessity and other modalities are structural, as Corfield suggests, they do not need to be inferred from particular facts, or to be arbitrarily posited.

The kind of necessity associated with structural determination is quite different from unconditional predestination. I want to affirm the first, and deny the second. Structural determination only applies within well-defined contexts, so it is bounded. If we step outside of the context where it applies, it no longer has force. (See also New Approaches to Modality; Free Will and Determinism.)

Leibniz is more familiar to me than Kripke, so when I hear “possible worlds”, I have tended to imagine complete alternate universes à la Leibniz. “Worlds”, however, could be read much more modestly as just referring to Corfield’s ranges of variation.

Form as a Unique Thing

Ever since Plato talked about Forms, philosophers have debated the status of so-called abstract entities. To my mind, referring to them as “entities” is already prejudicial. I like to read Plato himself in a way that minimizes existence claims, and instead focuses on what I think of as claims about importance. Importance as a criterion is practical in a Kantian sense — i.e., ultimately concerned with what we should do. As Aristotle might remind us, what really matters is getting the specific content of our abstractions right for each case, not the generic ontological status of those abstractions.

One of Plato’s main messages, still very relevant today, is that what he called Form is important. A big part of what makes Form important is that it is good to think with, and a key aspect of what makes Plato’s version good to think with is what logically follows from its characterization as something unique in a given case. (Aristotle’s version of form has different, more mixed strengths, including both a place for uniqueness and a place for polyvocality or multiple perspectives, making it simultaneously more supple and more difficult to formalize.) In principle, such uniqueness of things that nonetheless also have generality makes it possible to reason to conditionally necessary outcomes in a constructive way, i.e., without extra assumptions, as a geometer might. Necessity here just means that in the context of some given construction, only one result of a given type is possible. (This is actually already stronger than the sense Aristotle gave to “necessity”. Aristotle pragmatically allowed for defeasible empirical judgments that something “necessarily” follows from something else, whenever there is no known counter-example.)

In the early 20th century, Bertrand Russell developed a very influential theory of definite descriptions, which sparked another century-long debate. Among other things (here embracing an old principle of interpretation common in Latin scholastic logic), he analyzed definite descriptions as always implying existence claims.

British philosopher David Corfield argues for a new approach to formalizing definite descriptions that does not require existence claims or other assumptions, but only a kind of logical uniqueness of the types of the identity criteria of things. His book Modal Homotopy Type Theory: The Prospect of a New Logic for Philosophy, to which I recently devoted a very preliminary article, has significant new things to say about this sort of issue. Corfield argues inter alia that many and perhaps even all perceived limits of formalization are actually due to limits of the particular formalisms of first-order classical logic and set theory, which dominated in the 20th century. He thinks homotopy type theory (HoTT) has much to offer for a more adequate formal analysis of natural language, as well as in many other areas. Corfield also notes that most linguists already use some variant of lambda calculus (closer to HoTT), rather than first-order logic.

Using first-order logic to formalize natural language requires adding many explicit assumptions — including assumptions that various things “exist”. Corfield notes that ordinary language philosophers have questioned whether it is reasonable to suppose that so many extra assumptions are routinely involved in natural language use, and from there reached pessimistic conclusions about formalization. The vastly more expressive HoTT, on the other hand, allows formal representations to be built without additional assumptions in the representation. All context relevant to an inference can be expressed in terms of types. (This does not mean no assumptions are involved in the use of a representation, but rather only that the formal representation does not contain any explicit assumptions, as by contrast it necessarily would with first-order logic.)

A main reason for the major difference between first-order logic and HoTT with respect to assumptions is that first-order logic applies universal quantifications unconditionally (i.e., for all x, with x free or completely undefined), and then has to explicitly add assumptions to recover specificity and context. By contrast, type theories like HoTT apply quantifications only to delimited types, and thus build in specificity and context from the ground up. Using HoTT requires closer attention to criteria for identities of things and kinds of things.

Frege already had the idea that logical predicates are a kind of mathematical function. Mathematical functions are distinguished by invariantly returning a unique value for each given input. The truth functions used in classical logic are also a kind of mathematical function, but provide only minimal distinction into “true” and “false”. From a purely truth-functional point of view, all true propositions are equivalent, because we are only concerned with reference, and their only reference (as distinguished from Fregean sense) is to “true” as distinct from “false”. By contrast, contemporary type theories are grounded in inference rules, which are kinds of primitive function-like things that preserve many more distinctions.

In one section, Corfield discusses an HoTT-based inference rule for introduction of the definite article “the” in ordinary language, based on a property of many types called “contractibility” in HoTT. A contractible type is one that can be optionally taken as referring to a formally unique object that can be constructed in HoTT, and whose existence therefore does not need to be assumed. This should also apply at least to Platonic Forms, since for Plato one should always try to pick out the Form of something.

In HoTT, every variable has a type, and every type carries with it definite identity criteria, but the identity criteria for a given type may themselves have a type from anywhere in the HoTT hierarchy of type levels. In a given case, the type of the identity criteria for another type may be above the level of truth-functional propositions, like a set, groupoid, or higher groupoid; or below it, i.e., contractible to a unique object. This sort of contractibility into a single object might be taken as a contemporary formal criterion for a specification to behave like a Platonic Form, which seems to be an especially simple, bottom-level case, even simpler than a truth-valued “mere” proposition.

The HoTT hierarchy of type levels is synthetic and top-down rather than analytic and bottom-up, so everything that can be expressed on a lower level is also expressible on a higher level, but not necessarily vice versa. The lower levels represent technically “degenerate” — i.e., less general — cases, to which one cannot “compile down” in some instances. This might also be taken to anachronistically explain why Aristotle and others were ultimately not satisfied with Platonic Forms as a general basis for explanation. Importantly, this bottom, “object identity” level does seem to be adequate to account for the identity criteria of mathematical objects as instances of mathematical structures, but not everything is explainable in terms of object identities, which are even less expressive than mere truth values.

Traditionally, mathematicians have used the definite article “the” to refer to things that have multiple characterizations that are invariantly equivalent, such as “the” structure of something, when the structure can be equivalently characterized in different ways. From a first-order point of view, this has been traditionally apologized for as an “abuse of language” that is not formally justified. HoTT provides formal justification for the implicit mathematical intuition underpinning this generally accepted practice, by providing the capability to construct a unique object that is the contractible type of the equivalent characterizations.

With this in hand, it seems we won’t need to make any claims about the existence of structures, because from this point of view — unlike, e.g., that of set theory — mathematical talk is always already about structures.

This has important consequences for talk about structuralism, at least in the mathematical case, and perhaps by analogy beyond that. Corfield argues that anything that has contractible identity criteria (including all mathematical objects) just is some structure. He quotes major HoTT contributor Steve Awodey as concluding “mathematical objects simply are structures. Could there be a stronger formulation of structuralism?”

Thus no ontology or theory of being in the traditional (historically Scotist and Wolffian) sense is required in order to support talk about structures (or, I would argue, Forms in Plato’s sense). (In computer science, “ontology” has been redefined as an articulation of some world or domain into particular kinds, sorts, or types, where what is important is the particular classification scheme practically employed, rather than theoretical claims of real existence that go beyond experience. At least at a very high level, this actually comes closer than traditional “metaphysical” ontology did to Aristotle’s original practice of higher-order interpretation of experience.)

Corfield does not discuss Brandom at length, but his book’s index has more references to Brandom than to any other named individual, including the leaders in the HoTT field. All references in the text are positive. Corfield strongly identifies with the inferentialist aspect of Brandom’s thought. He expresses optimism about HoTT representation of Brandomian material inferences, and about the richness of Brandom’s work for type-theoretic development.

Corfield is manifestly more formally oriented than Brandom, and his work thus takes a different direction that does not include Brandom’s strong emphasis on normativity, or on the fundamental role of what I would call reasonable value judgments within material inference. From what I take to be an Aristotelian point of view, I greatly value both the inferentialist part of Brandom that Corfield wants to build on, and the normative pragmatic part that he passes by. I think Brandom’s idea about the priority of normative pragmatics is extremely important; but with that proviso, I still find Corfield’s work on the formal side very exciting.

In a footnote, Corfield also directs attention to Paul Redding’s recommendation that analytic readers of Hegel take seriously Hegel’s use of Aristotelian “term logic”. This is not incompatible with a Kantian and Brandomian emphasis on the priority of integral judgments. As I have pointed out before, the individual terms combined or separated in canonical Aristotelian propositions are themselves interpretable as judgments.

New Approaches to Modality

I periodically peek at the groundbreaking work on formal systems that is going on in homotopy type theory (HoTT), and in doing so just stumbled on an intriguing treatment of modal HoTT that seems much more philosophically promising to me than standard 20th century modal logic.

Types can be taken as formalizing major aspects of the Aristotelian notions of substance and form. Type theory — developed by Swedish philosopher Per Martin-Löf from early 20th century work by the British philosopher Bertrand Russell and the American mathematician Alonzo Church — is the most important thing in the theory of programming languages these days. It is both a higher-order constructive logic and an abstract functional programming language, and was originally developed as a foundation for constructive mathematics. Several variants of type theory have also been used in linguistics to analyze meaning in natural language.

Homotopy type theory combines this with category theory and the categorical logic pioneered by American mathematician William Lawvere, who was also first suggested a category-theory interpretation of Hegelian logic. HoTT interprets types as paths between topological spaces, higher-order paths between paths, and so on, in a hierarchy of levels that also subsumes classical logic and set theory. It is a leading alternative “foundation” or framework for mathematics, in the less epistemologically “foundationalist” spirit of previous proposals for categorical foundations. It is also a useful tool for higher mathematics and physics that includes an ultra-expressive logic, and has a fully computational interpretation.

There is a pretty readable new book on modal HoTT by British philosopher David Corfield, which also gives a nice introductory prose account of HoTT in general and type theory in general. (I confess I prefer pages of mostly prose — of which Corfield has a lot — to forests of symbolic notation.) Corfield offers modal HoTT as a better logic for philosophy and natural language analysis than standard 20th century first-order classical logic, because its greater expressiveness allows for much richer distinctions. He mentions Brandom several times, and says he thinks type theory can formally capture many of Brandom’s concerns, as I previously suggested. Based on admittedly elementary acquaintance with standard modal logic, I’ve had a degree of worry about Brandom’s use of modal constructs, and this may also help with that.

The worry has to do with a concept of necessity that occasionally sounds overly strong to my ear, and is related to my issues with necessity in Kant. I don’t like any universal quantification on untyped variables, let alone applied to all possible worlds, which is the signature move of standard modal logic. But it seems that adding types into the picture changes everything.

Before Corfield brought it to my attention, I was only dimly aware of the existence of modal type theory (nicely summarized in nLab). This apparently associates modality with the monads (little related to Leibnizian ones) that I use to encapsulate so-called effects in functional programming for my day job. Apparently William Lawvere already wrote about geometric modalities, in which the modal operator means something like “it is locally the case that”. This turns modality into a way of formalizing talk about context, which seems far more interesting than super-strong generalization. (See also Modality and Variation; Deontic Modality; Redding on Morals and Modality).

It also turns out Corfield is a principal contributor to the nLab page I previously reported finding, on Hegel’s logic as a modal type theory.

Independent of his discussion of modality, Corfield nicely builds on American programming language theorist Robert Harper’s notion of “computational trinitarianism”, which stresses a three-way isomorphism between constructive logic, programming languages, and mathematical category theory. The thesis is that any sound statement in any one of these fields should have a reasonable interpretation in both of the other two.

In working life, my own practical approach to software engineering puts a high value on a kind of reasoning inspired by a view of fancy type theory and category theory as extensions or enrichments of simple Aristotelian logic, which on its formal side was grounded in the composition of pairs of informally generated judgments of material consequence or material incompatibility. I find the history of these matters fascinating, and view category theory and type theory as a kind of vindication of Aristotle’s emphasis on composition (or what could be viewed as chained function application, or transitivity of implicit implication, since canonical Aristotelian propositions actually codify material inferences) as the single most important kind of formal operation in reasoning.