Essence and Identity

I'd like to feature today a post by infimitas at the IPS forum with the above title. For me it succinctly states the postmetaphysical approach so I'll let him speak in his words, following:


An essence refers to what something “is”.  This works by having a set of essential properties or qualities, which act as necessary and sufficient conditions for it to “be” that thing.   For example, I can say of a poodle that it “is” a dog.  This implies that the essential features of “dogness” (an abstract entity, so to speak) are instantiated in that particular (concrete) poodle.  I hope I’m not over-simplifying things here.  There’s a lot more that I could say that doesn’t seem all that relevant, so I’m leaving it out.


As a result of all this, I can say that cats “are not” dogs because they do not meet the necessary and sufficient conditions of “dogness”.  Note the bivalent (0 or 1; true or false; black or white) logic at play here – either something “is” that thing or it “is not” that thing.  Robert Anton Wilson wrote strongly against this kind of “either/or” logic, rightly I think (though perhaps too aggressively at times).  In any case, even if one distrusts RAW, in my opinion Wittgenstein had already effectively demonstrated years ago that there are no essences in the sense just described, and I further suggest that, taken to its logical extreme, his point also implies that we can’t use bivalent logic to describe objects without doing metaphysics.

To illustrate his anti-essentialist ideas, Wittgenstein used the example of games.  He pointed out that there is no essential description that fits all games – no fully consistent set of necessary and sufficient conditions that applies to all games.  This results in a sort of Goldilocks paradox, where any attempt at creating an essential definition either results in a standard too permissive, that lets too many things count as games, or a standard too restrictive, not letting in things that uncontroversially count as games.  For example, if we treat the quality of using balls as a necessary condition, then games that do not use balls are excluded.  If, however, we treat using balls as a sufficient condition, then anything that involves balls counts as a game.

Instead, Wittgenstein suggested that definitions function instead more like a sort of family resemblance model.  Imagine a family that has lots of tall members.  That is, most, if not all, have this quality.  Also, most, but not all, have black hair; and most, but not all, have large noses, etc.  So some might be tall and have black hair, but a normal nose; and some might be tall and have large noses, but not black hair, and so on, for all combinations.  The family resemblance, then, consists of a set of overlapping features, with no, one necessary and sufficient definition that works for everyone.

This generalises to all “objects”.  For example, a while ago, I saw a media programme about technology featuring a journalist trying to decide where computers begin and where they end.  He phrased the question as “what is a computer?”  But if there is no “computerness” essence—no essential definition of computers—then the question makes no sense.  “Is” an iPod a computer?  What about a Smartphone?  In effect, we can only say things like, “for legal purposes, that counts as a computer” or, perhaps, “I don’t consider that a computer”.

Unfortunately, these family resemblance models—nominal essences, if I can call them that—change with time and experience, according to needs or perspective, or across cultures (or subcultures... or even from individual to individual).  Wittgenstein used the rules of tennis as an example.  There is no rule stating that one can’t throw the ball 200’ in the air in order to serve.  But if someone started doing that, a new rule banning it might come into effect very quickly.  In this way, nominal essences can change; they do not have a fixed nature.  (I’ll give a real-life example shortly.)

The situation gets worse, however.  Not only do nominal essences have a fluid and imprecise nature, but we can’t even be absolutely sure we are all using words to mean the same things.  (I believe I am saying something similar to Derrida, where we can’t be totally sure that the same signs have the same referents.)  I don’t think this problem of communication need concern us too much, but it’ll take me too long to explain that point here.  Another post, perhaps.  Anyway, that example...

Kant suggested that two words with same meaning can be used synonymously.  He called this an analytic truth.  The classic example being: “All batchelors are unmarried men”.  But if the meanings of both “bachelor” and “unmarried man” refer to fluid, imprecise nominal essences, rather than rigid and fixed essences—like Platonic ideals, equally transparent to all—then we can’t really say that bachelors “are” unmarried men.  This point was made by the philosopher W.V.O. Quine.  Rigid essences don’t exist – or, perhaps more fairly, we have no right to insist that they exist.  To say that one knows the essence of something seems like bad metaphysics – an article of faith.

A concrete instance of this might help to flesh it out.  Under UK law, homosexuals cannot marry.  Therefore, using Kantian logic, if we know that George “is” gay, one could say that, by definition, George “is” a bachelor.  However, relatively recently, the law was amended to allow homosexual couples to register under “civil partnerships” (a legally recognised partnership, but not technically a marriage).  So if George “joins” with his partner under a civil partnership, “is” he still a bachelor?  One person might want to answer “yes” – a conservative, say, who wants so stick literally to the old definition, because the idea of gay relationships offends his sensibilities.  But a more liberal person might not consider George a bachelor anymore because he is in a legal partnership, which, although it might not be called a marriage by law, to that person, it still counts as a marriage in spirit.

So we can say neither that George “is” a bachelor nor that he “is not” a bachelor, because either description implies that there is a “bachelor” essence we have access to.  Really, we can only say something like “some people consider George a bachelor”.  That might sound silly to some people, but there are other ways in which using the “is of identity” results in dogmatic metaphysical proclamations.  Consider Wilber’s language (paraphrasing from memory, with italics added for emphasis: “... that is beautiful. It really is.”  But if we assume that no one knows the essence of beauty, we should instead say something like, “some people consider that beautiful”.  That adds context and forces us to admit perspective – it changes beauty from an object to a suobject.  (Or considering that socio-linguistic concepts are socially constructed, an inter-suobject.)

Another example, one that RAW used a lot: light “is” a wave and “is” a particle.  Saying instead something like “In some experimental conditions, light behaves as a wave; in other conditions, it behaves as a particle” avoids invoking the idea that we know the essence of light.  I asked a physicist friend today and he agreed with me that this sort of essence-agnostic language seems more appropriate for fundamental science.  Obviously, I know relatively little of physics myself, but it seems to me that if we treat scientific theories as just models rather than literal, metaphysical truths, then we acknowledge our own role in the perception of reality.  The world ceases to consist of “things” that we view objectively, and shifts to a more confusing (but more metaphysically honest) mass of (inter-)suobjects.

That shouldn’t be taken as license to annoy people.  I recommend a charitable reading of everyday language, e.g. we can assume that “the movie was great!” means the person liked the movie.  For metaphysically sensitive topics, however (or other things, if we feel up to the task), I suggest we drop the “is of identity” altogether.