A Step in the Direction of a Substantive Account of Identification

The goal of my previous post was to motivate the project of providing a theory of identification, as this concept occurs in UAP/UFO discourse. My goal in this post is to make some claims about (the) concept(s) of identification that appear to play a role in UAP/UFO discourse, and then to take a tentative step toward a theory (or account) of identification. As I see it, the four main claims that I’ll assert in sections 1-3 are, or should be, uncontroversial, in the following sense: they are, or should be, taken as desiderata for any theory (or account, or conceptualization) of identification. That is, any acceptable theory of identification should “fit with” each of these desiderata.

Some of what I say below might seem obvious. I certainly hope that at least some of it is; but I still think that it is worth saying and clearly explicating. Paying close attention to alleged obvious claims can be very fruitful in the long term. On the other hand, someone might reject this alleged obviousness. That too is fine: when you have a claim that one philosopher takes to be obvious but another takes to be incredible, that’s when you have the most interesting dialogue. I’ll start with what I take to be the most obvious theses, and proceed to what might seem less obvious.

1. The first desideratum that any acceptable theory of identification should accommodate is that whether something is identified can vary across time. At one time, something might not be identified, while at another time, when further information is available, it might easily be identified. Let us call this the Temporal Variability of Identification Thesis (TVIT). As I see it, TVIT can be accommodated by a theory of identification very simply, by relativizing identification to times (see below).

The second desideratum is closely related to the first, because the above quick explanation of the first requires that whether something is identified can change when further information is available. This exposes the fact that whether something is identified can vary across subjects (or agents), depending upon their epistemic situations. One birdwatcher might not be able to identify a particular bird because he is a half mile away from it, while a second birdwatcher easily identifies it because she is significantly closer to it. It is important to allow that the subjects/agents might be more than mere individuals: groups, or collective agents, can also identify things. Examples include: AARO, the scientific community, the UFO community, the public, the U.S. government, experiencers in general, and so on. Let us call this the Subject Variability of Identification Thesis (SVIT). As I see it, SVIT should be accommodated in a theory of identification by relativizing identification to subjects, in the same way it should be relativized to times.

Taking these first two desiderata together, then, a theory of identification should provide criteria for the application of the following phrase: ‘subject x identifies object y at time t’. Notice that this will result in fundamental identification claims that are fairly different from claims you often hear in UAP/UFO discourse. Most often the claims you hear are absolute (or unrelativized): ‘Fravor’s tic-tac cannot be identified’, ‘There were many sightings of unidentified objects in 1952’, etc. The goal, then, is to translate or interpret these ordinary claims into something that a theory of identification can process. Thus, the first might be interpreted as follows: ‘As of now, nobody is in a position to identify Fravor’s tic-tac’. And the second might be translated thusly: ‘In 1952, there were many sightings of things that, as of now, nobody is in a position to identify’.

(These are, of course, not the only ways to interpret these sentences. The claims are somewhat undecided about who is supposed to be doing the identifying. To see this, consider the alleged possibility that there is a UFO cover-up within the U.S. military-industrial complex. Just suppose, for a moment, that some conspiracy theory to this effect is indeed true. If so, then it is very plausible that there is someone who is currently in a position to identify Fravor’s tic-tac. If so, then the above translation of ‘Fravor’s tic-tac is unidentifiable’ is untrue. Is this intuitively the correct upshot?)

Much more can be said about the translation of UAP/UFO discourse into a paradigm that directly accommodates TVIT and SVIT, but I’ll move on.

2. I have listened to at least a couple of recorded talks by Kevin Knuth where he expresses the third desideratum that I have in mind. I recall that he expressed the thesis like this: identification comes in degrees. I forget Knuth’s exact example that he used to illustrate this, but a similar illustrative example might involve a birdwatcher who sees a small bird at a distance of 100 meters. At first he can only say that he knows it is a bird. Then he approaches closer, to 50 meters away, and looks through his binoculars, and he can say that he knows it is a hummingbird, but he can’t say exactly what species. He approaches closer, and once he is within 25 meters he is in a position to say that it is a Costa’s hummingbird, because he can see particular identifying features that distinguish it from other similar species. In some important sense, the birdwatcher can identify the hummingbird better as he approaches it; the hummingbird is more identified the closer he gets. Let us call this the Gradability of Identification Thesis (GIT).

(It is compelling to think that we might structure our theory of identification to accommodate GIT by building it in to our fundamental locution, as we did with TVIT and SVIT. An account along these lines might assert that a theory of identification should provide criteria for the application of the phrase, ‘subject x identifies object y at time t to degree d’. Such a theory would have to then explicate the structure of these “degrees of identification”: Can we represent the degrees as having the structure of the real numbers between 0 and 1, from completely unidentified to perfectly identified? Or should they have some other structure, open-ended on one or both ends, more closely corresponding to positive real numbers, negative real numbers, or all the real numbers? I ask these questions not in order to object to GIT, or this understanding of it, but rather because I really do not know what the correct answer should be.)

3. The final desideratum that I have in mind is that ‘identify’ and its cognates are ambiguous. I’ll call it the Ambiguity of Identification Thesis (AIT). Strictly speaking, ambiguity is a linguistic thesis, about the way a word is used in ordinary discourse. The easiest way to corroborate an ambiguity claim typically involves opening a dictionary (or, in the 21st century, typing the word into the Google search bar). We can do this with ‘identify’, and we find a few important uses (these from the American Heritage Dictionary). (A) Identification-as-recognition: “To establish or recognize the identity of; ascertain as a certain person or thing”; (B) Identification-as-classification: “To determine the taxonomic classification of (an organism)”; (C) Identification-as-characterization: “To ascertain as having a certain characteristic or feature”. Consult your linguistic intuition, and you can grasp these three different uses (although it is much more difficult to say whether or how they are interrelated).

Neither (A) nor (C) correspond to the concept of identification that concerns us. The mere fact that you cannot pick out a particular flying saucer from a suspect lineup should not imply that you cannot identify it; identification-as-recognition is not what we are getting at. On the other hand, the mere fact that you can characterize some flying saucer as having some feature or other (e.g., appearing to be made of some sort of dull silvery-gray metal) does not mean that you can identify it; identification-as-characterization is not what we are getting at, either.

So, by elimination, identification-as-classification would appear to be the crucial idea. On this conception, our fundamental locution ‘x identifies y at t’ can uniformly be interpreted as ‘x classifies y at t’. This view incorporates much of what we mean by ‘identify’ and its cognates in UAP/UFO discourse, but there are problems for the view.

4. To unearth the problems, we must first ask what it means to classify some particular thing. As I see it, classifying is an activity that is performed when a “classification system” is presupposed, and then an object is “situated within” that classification system. When our birdwatchers classify their object of study as a Costa’s hummingbird, they might presuppose whatever classification system is used in their favorite bird book, and they situate it within the system by indicating that it is a member of the kind represented on a certain page of that book (p. 230 in my edition of Sibley Birds West).

But not just any classification system will do. I might presuppose a classification system that partitions the world into kinds where the members of each kind are qualitatively indiscernible. Or (what may end up being the same thing) I might presuppose a classification system that creates a partitioning where each cell contains only one lonely object (each cell is the singleton of an object). How easy it will be to identify things while presupposing this system: just “situate” each object snugly within its own singleton — awfully hard to fail!

Presumably, then, on this view, when we aim to identify something, we must presuppose an explanatory classification system that is representative of our considered, organized scientific understanding of the world. There is a lot packed into that phrase, and I do not intend to unpack it all here. It is presumably very controversial what is the correct classification system that deserves that description. But I suppose the main idea is that there is some classification system that incorporates the core of all scientific understanding. We might call it our ‘considered worldview’ or ‘consensus reality’ (a term that is commonly used in the UFO community). We might go further and say that an ontology can be derived from this classification system — lions and tigers and bears are contained in it (because they are conceded by our core considered worldview to exist), but not unicorns and dragons and basilisks (because they are not so-conceded). Let us call this classification system ‘The System’.

This understanding of the classification-as-identity view seems to me to be fairly compelling. As I have explicated it, the view implies the following two claims:

NEC. Necessarily, if someone, x, identifies something, y, at a time, t, then x classifies y in The System at t.

SUF. Necessarily, if someone, x, classifies something, y, in The System at a time, t, then x identifies y at t.

While I find NEC and SUF initially compelling, I do not think they represent the best conception of identification as it appears in UAP/UFO discourse. This is because I believe that the three disambiguations of ‘identify’ that are contained in the American Heritage Dictionary (which I cited above) are not exhaustive of our concept of identification. I am unsure whether to say that this represents another use of the term; I do not know whether it is a true linguistic ambiguity. But I do know that it represents a different and important concept that at least deserves the name ‘identification’: (D) Identification-as-knowledge-of-essence: “To come to know what an object is; to learn of its essence”.

In the remainder of this post, I’ll present counterexamples against both NEC and SUF. In doing so, I hope to explicate more thoroughly the conception of identification that I have in mind.

5. Suppose that Henry the hermit lives in a cabin in the woods, and has done so his whole adult life. He has no contact with civilization. He gets no news from the outside world. He is totally isolated, and he has no memory of ever having been part of society. But Henry does have friends: the animals. In particular, he likes to feed the squirrels. He knows what they like to eat; he understands their life-cycles; he has even dissected them when they pass away. He has a deep understanding of squirrel-kind. Whenever Henry encounters some particular squirrel at a reasonable distance, he knows what it is, and so he identifies it quickly and easily. (Presumably, he does not use the term ‘squirrel’ — maybe he has developed his own language — but he identifies it nonetheless.) However, Henry cannot situate this squirrel within The System. He has no understanding of the public scientific enterprise that our civilization has created. He doesn’t even know that there is such a thing. Perhaps he has his own inchoate classification system, but it is nowhere near the sophistication and complexity of The System. Therefore, NEC is false.

6. Suppose that some great programmer has managed to build a program that incorporates all the nitty-gritty details of The System. As a result, for any information about some particular thing that you input, the program will do its best to situate that thing within The System — and when it is unable to determinately classify something, it outputs probabilities that the thing should be put in one part of The System or the other. Suppose further that Donald is an ignorant know-nothing, but he has access to this program via an app on his phone. Suppose finally that you observe some object behaving in a certain way, you compose a report of that behavior, and you give it to Donald. He plugs your report into the app, and it outputs a certainty that the object is a Costa’s hummingbird. In this situation, Donald has successfully classified the Costa’s: he has situated that particular hummingbird within The System for you. But he has not identified it, because he knows nothing about the hummingbird. He has merely performed the rote task of receiving your report and entering an input into his phone (if you like, we can presume that Donald doesn’t even know how to read — he just sees the letters that you wrote, and types them in). Therefore, SUF is false.

7. I intend these examples to demonstrate what I take to be the primary deficiency of the identification-as-classification view, which is encapsulated in NEC and SUF. That deficiency is that classification is not necessarily an epistemic activity that must involve epistemic agents. Indeed, I could have taken Donald out of the latter story, and replaced him with a robot. It would have been true to say that the robot classifies the Costa’s hummingbird, and equally true that it does not identify it.

I also intend these counterexamples to help elucidate what is currently my preferred view, which conceives identification as knowledge-of-essence. On this view, when you identify something, you are, roughly speaking, in a position to answer the question of what it is. The question of what something is cannot be answered merely by indicating some accidental characteristic of the thing; intuitively, to do that is to dodge the question, or not to answer it at all. In order to answer the question, you have to indicate that you know of some essential characteristic of the object. I believe that this view incorporates what is compelling about the identification-as-classification view. Often, when you know what something is, you are in a position to situate it within established scientific understanding (especially when you have a solid grasp of The System). And often, when you so-classify something, it is correct to say that you know what it is. So I accept the following revisions of NEC and SUF:

NEC*. Necessarily, if someone, x, identifies something, y, at a time, t, then x knows what y is at t.

SUF*. Necessarily, if someone, x, knows what something, y, is at a time, t, then x identifies y at t.

8. This post has gone on far too long. I’ll close with some notes about how I am inclined to flesh-out this view. These are very cursory thoughts, but I have reason to think that they would be fruitful avenues to pursue.

i. On essences. Talk of essences is highfalutin’ philosophy. Many good philosophers have shunned this sort of talk (think of Quine). I do not claim to be an expert in de re modality. I may have once had a legitimate claim to an "adequate understanding" of the topic, but I am at least a decade out of practice. In general, I am most inclined to accept something along the lines of the account of de re modality — and of essences, and the question of essentialism — that David Lewis espouses in “Counterpart Theory and Quantified Modal Logic” (in his Philosophical Papers, Volume I, pp. 26-46) and On the Plurality of Worlds (Ch. 4, esp. pp. 248-63). The following passage from the former reference should provide a rough understanding of where I am coming from. Lewis writes,

“I am by no means offering a wholehearted defense of ‘Aristotelian essentialism’. For the essences of things are settled only to the extent that the counterpart relation is, and the counterpart relation is not very settled at all. Like any relation of comparative overall similarity, it is subject to a great deal of indeterminacy (1) as to which respects of similarity and difference are to count at all, (2) as to the relative weights of the respects that do count, (3) as to the minimum standard of similarity that is required, and (4) as to the extent to which we eliminate candidates that are similar enough when they are beaten by competitors with stronger claims. […] Further, as with vagueness generally, the vagueness of the counterpart relation — and hence of essence and de re modality generally — may be subject to pragmatic pressures and differently resolved in different contexts. […] The true-hearted essentialist may well think me a false friend, a Quinean sceptic in essentialist’s clothing.” [Postscript to "Counterpart Theory and Quantified Modal Logic”, p. 42]

ii. On “knowledge-what”.  I am unfamiliar with any recent philosophical literature on knowledge-what, or on the semantics of ‘what-is-it?’ questions. Indeed, I am unfamiliar with any literature whatever on the semantics of knowledge-what questions, or really whether there is a significant body of literature. If there isn’t, then there should be. I am inclined to think that there must be some text from Aristotle where he discusses various interpretations of the question, “What is it?” In any case, it seems to me pretty clear that usually when we utter a question like this, what we are really getting at has something to do with essences, and so Aristotle should have said something about it.

iii. On the gradability of identification.  I am going to be bold and close with a strong claim: this view, encapsulated in NEC* and SUF*, can smoothly provide an explanation of GIT. Let us return to the example above in section 2, involving the birdwatcher who approaches the Costa’s hummingbird. At first, he identifies it as a bird, then he identifies it as a hummingbird, then he identifies it as a Costa’s. As he moves closer, he can better identify it. Question: What is it in virtue of which the bird is more well-identified as he approaches it? Answer: He knows more about what it is as he approaches it. So, on this view, the gradability of identification can be cashed out in terms of “knowing more” (which is something that is independently of philosophical interest, and I think there’s a philosophical literature on this).