PRH | Essay | 7.25 • Blank and the Categorical Skeleton

Blank and the Categorical Skeleton Blank is just Blank Aleksandar Perišić November 2025 Abstract Blank is often heard as a metaphor for “nothing”; in this note we argue that it can be treated as a precise structural object, on the same footing as the symbol 0in ordinary mathematics. Starting from the observation that 0is a dormant but fully specified carrier of potential—a kernel, a base point, a typed empty list—we push this role up one meta-level and formalize Blank as a zero-information pre-object. On the categorical side we introduce a simple “Blank-extended” skeleton consisting of a distinguished object I , a Blank object □ , and a one-way transfer arrow τ:□→I that encodes the irreversible act of choosing a first model. This makes explicit where the cost of “having a category at all” is paid, and how it relates to the blur paradigm in which we track information loss inside a fixed theory. The picture we propose is deliberately minimal: Blank is optional, but once adjoined it cleanly separates three layers—no model, bare skeleton, and blurred concrete theory—and turns the informal talk of “starting from nothing” into a small, categorical piece of data. 1 Blank and the categorical skeleton When we talk about Blank it is easy for it to sound like a loose metaphor: something that might or might not be there, a poetic way to talk about “nothing”. In this section I want to argue that Blank is no more mystical than the role already played by the symbol 0in ordinary mathematics, and that it can be attached to the categorical picture in a precise (but optional) way. It is well known that 0is one of the most potent objects ever shaped in mathematics. Historically it arrived late: many early civilizations either lacked a clear zero symbol or used it in a very restricted bookkeeping sense. Yet the real power of 0is not historical but structural. When we write 0, we are not referring to “nothing”; we are giving a dormant but fully specified object on which our theorems, functions, and projections can act. A few standard examples make this explicit: • In group theory, the kernel of a homomorphism is, again, a group. One often thinks of 0 as the neutral element or the entire kernel as the “zero layer” of the morphism: dormant structure which has not yet propagated outward, but is already fully defined. • In programming terms, 0is closer to a class definition than to an empty region of memory. A type definition or class declaration is a dormant skeleton: it says exactly what an instance could do, without activating any particular instance yet. • A typed list of length 0still carries information: it knows the type of its would-be elements. Likewise 0among the natural numbers is not a hole: it pins down the successor structure n7→ n+ 1 even before any counting begins. In all these cases 0is not an absence but a placeholder for potential, fully specified and yet, in itself, inactive. Blank, as developed in our earlier work (see the Essay on the □ operator), is simply 0pushed to the highest possible meta-level. There we treated □ as a zero–information ideal: an object 1 with no extractable bits at all; even mentioning it incurs an information cost, and any further structure we attribute to it immediately moves us away from the ideal. In the blur paradigm, Blank is essential: it is the marker for “no model bits spent yet”; every blur and every theory moves a small, explicit distance away from □. Categories, identity, and the hidden role of Blank From the point of view of category theory, Blank seems to live outside the formalism. The axioms of a category speak of objects, arrows, composition, and identity morphisms. They are written to be instantiated in many concrete settings (sets and functions, groups and homomorphisms, topological spaces and continuous maps), but the act of instantiating is left in the background. That act is where Blank hides. Conceptually, Blank lives in the pre-theoretic (one might even say pre-historic) phase of a theory: the stage before any particular objects or morphisms have been chosen, when we have not yet spent a single bit on a model. For the present discussion it is convenient to look at the simplest possible categorical skeleton: the walking object. This is the category with •a single object I, and •a single morphism idI:I→I, together with the usual identity and composition axioms (which here are trivial). This tiny category already encodes the pattern “there is a thing, and there is a do–nothing operation on it”. In the examples above we recognise I as a placeholder for 0(kernel object, empty list with type, base point of N ) and idI as the guarantee that acting on that dormant object does nothing yet: all actions we might care about are still to be specified. Notice that even this minimal categorical picture is already a model. It assumes we have decided on some universe of discourse in which I and idI make sense, and it assumes we are willing to apply this pattern in concrete categories later. Two crucial steps are therefore left implicit: 1. the step where we extract the categorical skeleton from some richer mathematical practice (groups, rings, graphs, . . . ), 2. the step where we apply that skeleton back to a particular theory. Standard category theory does not talk about either step explicitly. Objects and arrows are simply “there”, ready to be used. Blank, however, is exactly this ability to move between models: the latent capacity to form and apply the categorical pattern at all. At this point a natural objection appears. In category theory we already have a way to talk about transfer: functors and other morphisms between categories. Why are these not enough? The reason is that a functor F:D → C presupposes that both source and target categories D and Chave already been formed. It describes how to move structure once the theories exist; it does not describe the prior act of going from “no theory at all” to a first categorical skeleton. Blank is meant to model exactly this pre-theoretic step: starting from no fixed category and then paying the cost to obtain a first object Iand its identity morphism. Adjoining Blank as a one-way transfer arrow If we want to make this meta-layer visible, one simple device is to extend the walking object by aBlank object and a single transfer arrow. Formally, consider a small category C together with •a distinguished object I∈Ob(C)(our abstract “0” or “generic instance”), •a new object, which we denote by □, 2 •a single morphism τ:□−→ I. Definition 1 (Blank-extended category).ABlank-extended category is a tuple ( C,I,□, τ ) where C is a small category, I∈Ob ( C )is a distinguished object, □∈Ob ( C )is a “Blank object”, and τ:□→I is a morphism in C . The semantic reading is that □ carries no internal structure (“no theory chosen yet”) and τ represents the one-way transfer in which we commit to using I as our generic instance. The intended semantics are: •□ does not carry internal structure; it stands for “no theory chosen yet”, or “no model bits have been spent”. • The arrow τ is the transfer: it is the one-time act of committing to a specific categorical skeleton. Passing through τ means: we fix I and idI as the stage on which concrete objects and arrows will later live. • There is, in general, no arrow back from I to □ . This is the categorical analogue of the erasure principle: once we have collapsed all upper theories onto their shared categorical pattern, we cannot reconstruct the full upper theory from that pattern alone. This one-way nature is not a flaw; it is the whole point. In ordinary practice we constantly perform irreversible extractions: • From a concrete group homomorphism we extract its kernel. The kernel is a perfectly good subgroup, but from that subgroup alone we cannot recover the original map. Information has been erased. • From a rich combinatorial or analytic situation we extract a category of objects and morphisms. The category captures the behavioural pattern but forgets the ambient details. Again, information has been erased. Blank is the limiting case of this erasure: if we erase all concrete theories, what remains is not a particular structure but the bare ability to form structures. In the Blank-extended picture this is symbolised by □ and the arrow τ:□→I . We do not claim that □ is the zero–information ideal from the □ –operator essay; rather, it is a categorical avatar that remembers where we pay the cost of choosing a model. Why this matters for blur Why bother adding such an invisible layer? Because blur lives exactly here. Blur is about controlling how much information we spend when we move from one model to another and when we tighten or loosen the lens. Without a place in the formalism where this transfer can be located, we risk talking as if model choice and application had zero cost. In the Blank-extended category, that cost is concentrated in the arrow τ:□→I: • every time we instantiate the categorical skeleton in a concrete theory (groups, numbers, graphs), we are conceptually traversing a copy of τ and spending a few bits of information to say “this is the object Ihere”, • every time we blur or unblur a theory, we are effectively changing how much is forgotten along such transfers. The impossibility of going back—no canonical arrow I→□ —is a reminder that precision cannot be arbitrarily increased for free. Once we have committed to a model, blur tells us exactly how much we can safely ignore; Blank tells us that we cannot rewind the commitment itself. 3 2 Relation to blur and further remarks In the companion article on blur categories (“A Category of Blur and the Grand Lemma”), we work entirely inside a fixed base category C : objects there are already proposition spaces or state spaces, and a blur structure is given by a family BX ( ε )of blurred avatars, together with reading maps and exhaustion at infinity. In that setting blur is a graded mechanism for controlling how much information is retained or forgotten within a chosen theory. It is natural to ask how the Blank operator defined here fits that picture. The answer is that Blank lives strictly one step earlier. The object □ and the arrow τ:□→I model the move from “no fixed theory at all” to the first categorical skeleton ( I, idI )on top of which a concrete category C and its blur structure can later be built. All the bits of information that define C and its blur are, conceptually, added on top of this pre-object. From this perspective, categorical blur can be seen as an expansion of τ . The single irreversible transfer τ:□→I is refined into a whole graded family of blur scales ε∈I and corresponding blurred objects BX ( ε ), each describing a partially realised, partially forgotten version of the same underlying model. The Blank article isolates the thin layer where we pass from no model to a first skeleton; the blur article describes, in detail, how information then flows and decays once that skeleton has been instantiated in a concrete category. In this sense the two viewpoints are complementary. Blank states that “being able to model” can itself be treated as a structural object, not just an informal background assumption. Blur then shows how, once a model exists, changing the resolution at which we query it has a precise categorical meaning. Together they justify the claim that blur and Blank are at least as structural and abstract as the usual categorical notions, and not merely a philosophical veneer on top of standard mathematics. References [1] A. Perišić, A Category of Blur and the Grand Lemma, Zenodo, 2025. 4