This series of essays has been rewritten from a Master's thesis defended at Temple University in 1970.

The Philosophical Foundations of Human Cognition (Section II)

©2004 Edward G. Rozycki

Part 3 : Recognition and Knowing


edited 7/26/15


(Section I) --------------



knowa and knowe

The Truth Condition

Enabling a Knowledge Claim


Epistemically Fundamental Verbs

Direct Reference

Epistemic Substitutes and Epistemic Sources

The Possibility of Illusion


Section Endnotes



(Section II) ------------


Contrasting 'recognize' and 'identify' Interesting Ambiguities

Two Senses of Recognize Performance vs Recall.

Recognition and Recall
The role of memory

The Primacy of Recognition
A justification for knowledge claims

Conflicts of Recognition and Identification Which one trumps?

Nelson Goodman's Languages of Art. Recognition-identification conflict.

Is Cognition Recognition?
Was Plato right?

Section Endnotes

(Section III) ------------


Confrontable Characteristics

Recognition Equivalence

Goodman's Identity Theory

Systematizing the r/e-class

Goodman's Theory Continued

Knowing What a Thing Looks Like


Dimension of Individuation

(Section IV) ------------

PART 4: Continued

Multi-leveled Individuation

NIE's and Classes

Physical Objects as NIE's

The Act as NIE

Following a Rule

Recognizing Reconsidered

Goodman's Theory -- Conclusion

The Application of Concepts

To Conclude



Contrasting 'identify' and 'recognize' (Interesting ambiguities)

If one identifies some particular object X as a Y, one is justified, on the basis of the identification, in claiming to know that X is a Y. We could say either "He identified X as a Y" or "He identified X to be a Y". Similarly, if one recognizes X to be a Y, such recognition counts as an identification, justifying the claim to know that X is a Y. In general, with the exception of merely recognizing discussed below, to recognize X as a Y is to identify it as a Y, justifying the claim to know that X is a Y.

By contrast, one may identify X as a Y even though one can not recognize X as a Y, as when one identifies an object decomposed beyond recognizability.

One may identify X in X's absence, But X must be present to be recognized. This suggests that certain characteristics used to identify X may differ in kind from those by which X is recognized. Let us distinguish between derivate characteristics, e.g. footprints, odors, signatures, by which a particular may be identified in its absence, and immediate characteristics, e.g. appearance, manner of speaking, which can be identified as characteristics of the particular only in the presence of the particular, and in terms of which we may recognize the particular.

Now, using derivate characteristics to identify presumes a normal history of derivation and the sufficiency of the characteristics to effect identification to the desired degree. For example, if I would identify a person by a footprint left in the snow, I must presume  or establish  that it is indeed a footprint  not a happenstance configuration caused by, say, some animal  and that it was brought about by someone's stepping in the snow;  not by someone else's dropping a boot just to leave a track. The footprint must be such as to individuate a particular person from among others who wear boots and could have left the footprint.

By contrast there is no presumption of normalcy of derivation in recognizing a person by his appearance. If I recognize John by his nose, I recognize the nose and since any nose on John's face is John's, I recognize him thereby. Of course there is the usual presumption that illusion is not taking place, but that is common to any identification. The big difference between recognition and identification is the connection between what one is confronted with, and the knowledge claim that is made vis a vis that confrontation. The following formulations are invoke our normal ways of making judgments about our practices of identification:

1. You can ask me to identify someone by his gait. You cannot ask me to recognize him by his gait. This suggests that it may be proper to speak of identificational criteria, but not recognitional criteria.

2. One can misidentify, but not misrecognize something. A failure in recognition is not a false recognition; it is not a recognition.

3. An identification can be partial; not so a recognition.

4. One can recognize and still misidentify; but only if a previous misidentification has occurred.

According to Webster's Third International Dictionary one can identify with X -- which I will not treat here -- or one can identify X to be something. Specifically (my analysis, not Webster's)  one can:

a. identify X to be a Y

b identify X to be Z, 'Z' a proper noun,

but not merely

c. identify X.

Such identification justifies the knowledge claims, respectively, that

a. X is a Y, and

b. X is Z.

When one identifies X to be Z, one individuates X, distinguishing it from others of the same species. 18

One can recognize X as a Y, or as Z and in terms of this recognition justify the claim to know that X is a Y, or Z. But unlike with 'identify', one may just merely recognize X.

Two Senses of Recognize (Performative vs Recall)

'Recognize' has two senses: a performative sense --'acknowledge' -- which I will indicate with recognize a ; and a knowledge-recall sense (Webster: recognize -- to recall previous knowledge of), which I will indicate with recognize r . One may recognize a John to be the King although John was heretofore a peasant. But to recognize r John as the King, John must have been king previously. To recognize a X as Z is to identify, i.e. acknowledge the identification of, X henceforth as Z. But to recognize r X as Z presumes that X was previously identified as Z.

Suppose X is a particular of a unique kind, or that we haven't the vocabulary with which to classify X. Suppose also that we recall on encountering X that we had encountered this same thing previously. We could say that we merely recognized X -- i.e. "merely recognizedr ," "merely recognizea" is not possible here
-- as the same unidentifiable object we had come across earlier. X may be unidentifiable although certainly recognizable.

Recognition and Recall (The role of memory)

If we remember the way a particular, M, looked and the X we are confronted with now looks the same  to the best of our memory, does this ipso facto establish that X is M? Consider that identity is not anything we can discern in the looks of things, otherwise we couldn't be fooled by imposters or counterfeits.

And what is this, our memory? A recall, a re-experience -- in some sense -- of what? A dream, an hallucination, an actual past event? The memory does not come labeled so as to indicate its source. And if you are sure your memory is not of a dream but of reality, is that any assurance for me? (Perhaps; if I think that your memory is good.)

If you claim to recognize X as Z what are you recalling? Not that X is Z since X is being encountered . The way Z looked ? Yes, to the best of your memory. But what establishes the identity of X as Z? Is it not in the fact that you treat X as Z and justify such treatment by the fact that you have certain memories? You presume of course that X is not some different object, indistinguishable in appearance from Z, which has been substituted for Z.

I will summarize our considerations in this manner:

we can analyze 'recognizer' as

a. 'recognizea' justified by recall,

b. presuming no indistinguishable replica has been substituted for that recognized, etc.

The words 'recall' and 'memory' are epistemically loaded. Memory is our common experience of deja vu. The epistemic validity of such an experience can be established only in the context of certain presumptions and given the availability of certain justifications.

Memory is a capacity and it may or may not function well. It is not per se a source of knowledge. If I claim to recognizer X to be a Y, I am confronted with X, recognize X as a Y, justifying such recognition, if necessary, on the basis of my memory of what a Y looked like. I enable the claim with the presumption that the appearance of Y's hasn't changed since my previous encounter with one. If I claim to recognize X as Z, I am confronted with X, recognizea X to be Z, and justify such recognitionaon the basis of a memory I have of what Z looked (smelled, sounded, etc.) like. My presumption that no indistinguishable replica of Z has been substituted enables my claim.

What about 'recognizer' as contrasted with 'claiming to recognizer'?

Unless I can establish the claim to recognizer X, the epistemic validity of a nonclaimed recognitionr is a presumption. Indeed, it is only in virtue of the presumption that my memory serves me well that I can designate my experience as recognitionr at all.

We might at this point venture to offer a parallel between 'recognitionr ' and 'claim to recognitionr' to 'knowe' and 'claim to knowe'. Just as there is no faculty that we possess in virtue of which we recognize, so is there no knowledge faculty. Recognitionr is object- or event-identity attribution justified by memory. Knowledgee is  -- to some limited extent (the notion is too complex to warrant a stronger generalization) -- predicate attribution justified in terms of epistemically fundamental "activities" and/or the faculties from which they derive. The common pattern here is : A performance justified by recourse to some faculty in the context of certain presumptions.

The Primacy of Recognition (A justification for knowledge claims)

The analyses to follow are independent of the conclusions stated in the immediately preceding section, which are offered as speculation along a particular line of thought rather than a carefully laid out foundation for the rest of my argument.

Of recognition and identification, which is conceptually more fundamental to an account of knowing?

Recognition, since:

a. recognition implies identification, but not the converse,

b. identification presumes recognition.

The sole exception to the generalization that recognition implies identification is mere recognition. One may in fact say, "I recognize it but I can't identify it." This is the situation depicted above where something recognizable was not identifiable. Upon recognizing something a person is not, strictly speaking, unable to make identificatory remarks about the object recognized. But such remarks may be uninformative to the listener below the level tolerated in a social situation.

For example, if you and I are confronted by some kind of beast, it would hardly do for me to offer as an identification, "That is the beast that now confronts us", although logically speaking I have classified it in contrast to all those beasts not present. Our logical requirements for an identification may be merely a predication which bisects our universe of discourse, but much more is demanded in the context of a natural language and the social setting of its users. Again, supposing that I recognized the beast I could say in virtue of this recognition , "I saw this one before." In a sense I have identified the beast, namely as "the beast which I previously encountered". This again won't do, for it is, according to traditional practice, no identification.

We may say then that with the exception of mere recognition , recognition implies identification and consequentially justifies per se certain knowledge claims. We will consider now an argument that an identification presumes a recognition.

If a knowledge claim is justified by an identification, we may ask of the claimant, "How did you identify it?" (We cannot in this context ask,"How did you recognize it?19). Suppose the answer is, "I saw that it had characteristic x, and having x the criterion for being (say)M." But we then ask the claimant how he identified x to be x. We can see that somewhere the claimant must run out of criteria. Somewhere he will have to say he recognizes certain characteristics when he sees them, because the alternative is an infinite regress. 20

Conflicts of Recognition and Identification (Which one trumps?)

One can misidentify but not misrecognize. I may say of someone, "I recognize him. He's the Duke of York." That I am wrong about his being the Duke of York does not mean that I don't recognizer him. I recognizer him as the same person whom, on a previous occasion, I had wrongly identified as the Duke of York. But then again, it may be the case that this is a man who looks very much like the Duke of York, whom I do know. If so, I retract my claim to have recognizedr him.

There is one other possibility which might bring me to retract my claim to recognizer someone, although it matters not here whether the identification is correct. If it can be established that I could not possibly have memories of the person be they through personal confrontation or pictures, etc. I might just chalk up the experience as a freak, a deja vu, having no epistemic significance. Dismissing this last possibility to simplify generalization we may say that

There are but two resolutions for a recognition-identification conflict:

a. Recognitionr is maintained as mere recognitionr and the identification is conceded as incorrect.

b. The claim to have recognizedr is withdrawn.

In extremely unusual circumstances we may decide that our criteria of identification are insufficient, and maintain the recognition and the identification consequent of it against the conflicting identification. For example, there might arise circumstances wherein we accept a recognition and its consequent identification of a person despite conflict with an identification established by fingerprint comparisons. If we were willing to concede that his fingerprints changed, the Harry of yesterday is the Harry of today. Otherwise we face an imposter.

Nelson Goodman's Languages of Art

From here on  -- in installments -- we will be considering an identification theory expounded in Goodman's Languages of Art. Goodman's theory runs completely counter to the one propounded in this paper. I will charge that Goodman completely misconceives the relationship between 'recognize' and 'identify'. Now what I offer here -- based on traditions of educated linguistic judgment -- is an analysis of these words and their interrelations. At the very foundation of his theory -- the identification of replicas or instances of atomic characters in a notation system -- Goodman relies on this tradition of analysis. But because of the criterion by which he would identify musical works he is forced counterintuitive conclusions, e.g. performances of a work that have mistakes cannot be allowed to count as performances of the work, and in effect destroys the theoretical base that he started from.

Goodman's criterion for judging a particular performance to be of a work W is that the performance comply with the score of W, i.e. an accurate transcription into musical notation of what is actually played must be identical with the score of work W. Goodman's criterion derives entirely and solely from the musically notatable characteristics of the performance. We will examine this criterion in a later section but right now consider the consequences of a recognition-identification conflict assuming the validity of Goodman's theory.

According to Goodman, we must reject a performance,p, as a performance of a work,W, if the derivate score of p, i.e. the score deriving from an accurate transcription of all the musically notatable characteristics of the performance, is not identical with the score -- antecedently identified -- of W. Now, suppose we recognize a piece being played as the Emperor Concerto, however, the performer makes a mistake. We may not -- according to Goodman --identify the piece as the Emperor Concerto since the derivate score of the performance is not identical with the score of the Emperor Concerto.

We have here a recognition-identification conflict. How do we resolve it? Do we withdraw our claim to have recognized the piece being played? On what grounds? We know the Emperor Concerto is what we have just heard an "imposture" or "forgery" of it? Is it a different work? Our memory serves us here to absolutely no avail. That we classify what we have just heard performed as either an imposture or a forgery or a different work or even a "nonwork"  disestablishes the epistemic relevance of recognitionr.We can never in the course of a performance identify what is being performed. It might be conceded that we do recognize the piece, but we merely recognize it; no identification is consequent to the recognition. Again, recognitionr is epistemically irrelevant.

Another interesting problem is this. A period of silence is musically notatable as a rest. Thus for any given performance we may get different derivate scores differing in the number of bars of rest preceding and following the notes actually played. Unless some reduction rules are introduced the identity of the performed piece cannot be established. When does the transcriber of the performance start transcribing? When does he stop? Why? We will return to Goodman's theory in the next part.

Is Cognition Recognition? (Was Plato right?)

With the exception of mere recognition, recognition implies identification. Therefore we can justify our claims to know what things are by recognition. But any claim to know what a thing is, i.e. an identification, must ultimately rest on recognition, which relates aspects of our immediate encounter with recall of that previously known.21 So far as knowing what things are, the Platonic slogan identifying cognition with recognition seems to ring true. In the Meno Plato gives terminus to the obvious regress by postulating that all knowledge, which has been accumulating through past lives, resides in the soul.  Learning is merely the eliciting of that which is already possessed. We needn't resort to such a maneuver, however, to answer the following argument.

Plato's argument is this: If I recognize, I recall knowledge. But how did I come by this recalled knowledge, i.e. how did I come to know previously what I now recall? By recognition, that is by recall of even anteriorly acquired knowledge. The argument leads to infinite regress. This is absurd. So there must be a beginning, perhaps in anterior lives, where we learned what we recall. (Are we forced to concede plausibility to Plato's argument?)

What Plato's argument does is to ignore the distinction between performative and knowledge-recall recognition. We extricate ourselves from the regress by hypothesizing a performative recognition, recognitiona, somewhere back in the chain. Performative recognition per se does not require any prior truths. If one recognizesa X to be Y, this is not because X is known to be Y, rather X is known (henceforth, from the time of instruction) to be Y because one recognizesa X to be Y.

The child learns that a certain object M is a chair, i.e. he recognizesa M on that occasion to be a chair. The next time he encounters M he recognizesr N to be a chair  given he has a memory and has learned that such memory experiences can serve to justify identity attributions.

But can't a child be mistaken and recognizea some N as a chair when it is really a table? Does not this indicate a prior standard?

A child can wrongly recognizea a table to be a chair. And the standard is not prior in the sense that some knowledge is logically prior to recognitiona. The child is wrong in the sense that his performance is inappropriate in the context of the usage of the speech community he is in. "We don't call that a chair!" is sufficient justification to withdraw recognitiona of N as a chair. That the child ought to withdraw such recognition under the circumstances is a presumption that maintains community.

Each speech community is definable in terms of a particular set of performative recognitions which provides members of the community with the foundation for its knowledge of things. What a thing is, is relative to any particular recognition-set, but within each set and in terms of it, judgments of correctness can be made.

Performative recognition, recognitiona, of certain X's as certain Y's is constitutive 22 of the semantic schema employed in and definitive of a given speech community. We teach a child what 'red' is not by having him attend to and notice some characteristic all red things or most red things  have in common; but by having him acknowledge in particular cases that this is red. ( I will discuss concept formation in the next part.)

The capacity to merely recognizer seems to be possessed by most forms of higher life. Mere recognition is the recognition of some particular as the same from one occasion to another. But by an analysis done above "recognizer" was an object identity-attribution justified in terms of memory. It would seem, since it is our common experience that dumb animals can recognize particulars, that the notion of acknowledgement needs some analysis, since while we may by analogy attribute memory to animals, what it is to acknowledge and whether this presumes intention on the part of one who acknowledges is important to know. In the last part of this paper I offer a clarification of what it is to follow a rule. I will return to 'acknowledge' then.


18 Webster, page 11

19"How did you recognize her underneath all that make up?" We are not expected to tell how.

20 Call an identification of x as (a) Y by characteristic c a C-identification. c is presumed sufficient to individuate Y's from non-Y's. Now either c confronts us and we recognize c or is in turn identified by its own characteristic m, which confronts us and is recognized by us or ... etc. ad infinitum . Note that if m is sufficient to identify c, it is sufficient to identify X as a Y and is superfluous.

21 In recognizingr x to be a Y, we confront X, recognize a X to be a Y justifying our performance by our recall, i.e. our claim to recall how a Y confronts one, i.e. the way it looks, feels, smells, etc., and certain presumptions.

In recognizing X to be Z, we confront X, recognize X as Z, justifying it by our recall of Z and certain presumptions.

In recognizing X to be a Y (or a Z) by c, a characteristic, we confront X and c , recognize c to be c justifying this by recall of and certain presumptions, prominant among which is the presumption of the sufficiency of c to establish the identity of a Y (or Z).

22 There are two ways of doing something incorrectly and distinguishing between them is a matter of assessing ability and intention. If a known English monolingual utters certain sounds, his performance, presuming intention on his part. may be judged incorrect with respect to the performance taken to be normally appropriate in his situation. However, for a multilingual person, such judgement may be withheld since he may not -- so to speak --   be still in the game. But there is a kind of speech ethnocentrism whereby not playing the game warrants a charge of incorrectness. Using the locution "is constitutive of" indicates that the kind of correctness involved in a "mistaken" performative-recognition is this latter kind.

If a, b, and c. are rules constitutive of game A, I may make a mistake by following b. where I should have followed c. or  ethnocentrically judged  by following d. I may be playing my own game, but in terms of A, I am incorrect.