Home - Wittgenstein's Logic of Language | Bibliography

Family Likeness - Criticism of Wittgenstein's Simile

This discussion follows Wittgenstein's investigation of games and family resemblances, where he presents his resolution of the problem of the meaning of a common name. He finds that common names often don't name a common nature (essence) but that there are instead similarities between bearers of a common name. These similarities he compared to the facial similarities in a human family, although he does not say how one family is to be distinguished from another. [Asylums of ignorance -- the limits of Wittgenstein.]

Outline of this page ...

Background: These notes belong to logic of language (which is Wittgenstein's expression from the TLP, but here used as my jargon: How is sense (meaning) to be distinguished from nonsense in the language used to discuss philosophical problems?)

Wittgenstein's Metaphor "Family Likenesses"

Query: objections to Wittgenstein's family resemblances.

What is the point of this metaphor -- and that is what it is: a metaphor, not a metaphysical theory --? If we wanted to define the word 'game' e.g. we might ask for a Socratic definition of that word: (1) what do all members of the class named 'games' have in common, and (2) what distinguishes this class from all others? An absolute definition would serve to tell us whether anything is or is not a game. But what if when we look at games we discover that games do not have a common nature or essence? And in fact "if you look at [games] you will not see something that is common to all, but similarities, relationships, and a whole series of them at that" instead (PI § 66).

Before rushing to the question of whether the metaphor "family resemblances" is apt (i.e. in what way this comparison does or does not do the work Wittgenstein intended it to do), we must ask about the meaning of Wittgenstein's concept 'family resemblances' (in this instance illustrated by the concept 'games').

Description versus explanation (theory)

Why is "family likenesses" not a theory? Because it does not explain how it is possible ("concept-formation") for us to know -- if 'know' applies to this case -- what to call or not to call a game. Family likenesses is instead only a description of the facts we find if we look at games e.g.

"This and similar things are called 'games'. And do we know any more about it ourselves? Is it only other people we cannot tell exactly what a game is?" (PI § 69) [But every explanation of the meaning of the word 'game' I can give myself I can give you too. (cf. ibid. § 210)]

But that leaves us in the condition of Socrates when he asks for a standard in Plato's Euthyphro, and does not find one. And therefore we have two possibilities -- either we can stay with the facts in plain view as Wittgenstein does, or we can try to invent a philosophical-metaphysical theory to explain or account for the facts. We can suggest an invisible explanation for what is visible. Examples of such theories are the "theory of abstraction" and Plato's "theory of Forms". Neither of these -- nor any other such theory (which, like all metaphysical theories, may be consistent with the facts of experience but cannot be falsified by the facts of experience) -- was, however, what Socrates was asking for. And Wittgenstein also did not want such theories (for which reason Bertrand Russell attacked his philosophy):

We must do away with all explanation, and description alone must take its place. (ibid. § 109)

That is Wittgenstein's program or project in his later work. It may be what we want from philosophy or it may not be. But even if what we want is a philosophical theory, we should nonetheless recognize that "family resemblances" is not one. It explains nothing. It is a metaphor: the way members of a class (e.g. games) resemble one another is like the way members of a human family resemble one another. And now we must look at exactly what comparison Wittgenstein wished to make, because he was not talking about biology.

Note: Wittgenstein's own English expression was 'family likenesses' (see The Blue Book below) rather than Anscombe's 'family resemblances'. And that expression, because it suggests both alike and dis-alike -- i.e. making comparisons of both types -- may be clearer in that respect. Family "alikenesses".

But there are many meanings of the word 'theory' (Comparison versus theory)

Query: Ludwig Wittgenstein, family resemblance theory.

Would there be a reason for calling a comparison (a metaphor, a simile) -- a theory? What would that mean, the word 'theory', in this context? Normally by 'theory' we mean 'an hypothesis that summarizes or makes a model of the raw data (the facts) of the case'. And then we ask: does the hypothesis stand up to the tests of reason and experience? Could we say that about Wittgenstein's simile 'family likenesses', that it states an hypothesis? How do you verify a simile? Well, either it is valid -- i.e. either it is possible to compare A with B in some specified way -- or it is not (the logic of comparison).

However, what does the word 'theory' add here? Is not the word 'comparison' enough by itself? Is the possibility of making this comparison denied by anyone on the basis of the facts in plain view -- i.e. because of what we find when we look at our language (rather than what someone imagines we ought to find, but don't)?

What is to be denied is that Wittgenstein specified a way to distinguish between different families -- and therefore the addition of the word 'family' may contribute more obscurity than clarity. Whether or not there is a more apt name (title) for this comparison (simile), I don't know. However, as the comparison now stands it cannot be used to define a word -- i.e. there is no such thing as "a definition by family resemblances" -- but only to describe a word's grammar as not being of the type where there is one feature common to everything that we call by that word.

Writing contrary things

When [the director of the Rockefeller Foundation] suggested that money be used to print any papers that [Wittgenstein] might have ... W. said: "But see, I write one sentence, and then I write another -- just the opposite. And which shall stand?" (Bouwsma, 11 January 1951, Oxford, Wittgenstein: Conversations, 1949-1951 (1986), p. 73)

And this is what happens here, for I had written just the opposite elsewhere.

And then later, that the expression 'family resemblances" says nothing more than the word 'similarities' alone would. But maybe logic would have a use for the expression "definition by similarity".

Query: fundamental concepts of Plato's Phaedo.

The word 'concepts' here seems to mean 'ideas' or 'notions', but it might also mean 'theories'. Both 'concept' and 'theory' are, as commonly used, concepts without strict borders ("fluid concepts"), often equivocal or unclear in intended meaning, "maid-of-all-work" words (That is what Guthrie calls the Greek word 'logos', our 'logic'). (Is the word 'theory' an important word?)

"Family Resemblances" - What doesn't it mean? (What does it exclude?)

I can think of no better expression to characterize these similarities than 'family resemblances'; for the various resemblances between members of a family: build, features, color of eyes, gait, temperament, etc. etc. overlap and criss-cross in the same way. -- And I shall say: games form a family. (PI § 67)

What does this simile make clearer? It is not obvious what Wittgenstein meant by it. If anything 'family resemblances' suggests an essence -- i.e. common ancestors; -- but that was not how Wittgenstein intended his metaphor to be applied. [The role of the word 'family' in this metaphor, simile or comparison.] If someone says "A is like B", we do not know what that person means until he tells us in exactly what way A is like [resembles] B. A metaphor is a comparison; it is not a statement of identity. 'A is like B' does not mean 'A is [identical to] B'.

But "what the sign slurs over, the use makes clear", and we can give examples to show Wittgenstein's use of this "sign" (in Wittgenstein's jargon: ink marks, spoken sounds, the purely physical part of language, in contrast to its use or meaning).

"Family Likeness" (The Blue Book)

The following statements are from the book Wittgenstein dictated to his students in 1933-1934.

We are inclined to think that there must be something in common to all games, say, and that this common property is the justification for applying the general term 'game' to the various games; whereas games form a family the members of which have family likenesses. Some of them have the same nose, others the same eyebrows and others again the same way of walking; and these likenesses overlap. The idea of a general concept being a common property of its particular instances connects up with other primitive, too simple, ideas of the structure of language. (BB p. 17)

There are an endless variety of actions and words, having a family likeness to each other, which we call ... (ibid. p. 33)

Here the expression "structure [i.e. form] of language" is an example of what Wittgenstein compared to the old shell still clinging to the newly emerged baby chick -- i.e. it says something new but still has egg-shell from the old view attached to it, the old notion clinging to a new idea (CV p. 44, a remark from 1944 or later). Because what Wittgenstein is talking about here is not a too simple idea about the form of language -- but instead a too simple idea about the meaning of language. (In Wittgenstein's later "logic of language", meaning is not a matter of form but of use.) Also, "primitive, too simple, ideas" is hardly a just characterization of Socratic definition as Aristotle describes it, nor of Plato's search for common natures as the meaning of common names.

We could easily describe primitive language-games showing how a tribe might begin by grouping together all their activities which made use of balls under the common-name 'game'. And then we can build onto this that some of these games played with balls also made use of nets. And in this way we can picture that the tribe added badminton to their list of activities called 'games', although badminton is not played with a ball. And then we can ask: what do all these activities that the tribe calls 'games' have in common now? And the answer naturally is: nothing. [That is an example of a fictitious natural history and of Wittgenstein's metaphor of "the ancient city" of our language.]

Suppose it were said that what all games have in common is that they are done for recreation as opposed to being work or religion e.g. But do we call travel "to see the sights" or reading popular fiction or going to the movies 'games'? Besides which, perhaps the tribe puts the loser at badminton to death (like a gladiator), or the winner might be sacrificed to the tribe's god; in which cases we might not want to compare this tribe's games to our recreations nor to distinguish its games from its religion (cf. the Olympic Games of the ancient Greeks). Also, we can imagine that the tribe called trawler-fishing a 'game' because it is done with a net, although it is work, not recreation, for the fishermen who do it.

This metaphor also appears in Wittgenstein's "Remarks on Frazer's Golden Bough", where he suggests drawing lines to connect the various features that faces of a family share: not all faces will share the same features. It is also important, I believe, to note that: facial resemblances are usually not exact: identical twins are not the usual case. So the resemblances may be blurry as well as sharp.

"All the members of that family have the same shaped noses." -- A feature common to all was not what Wittgenstein had in mind (except perhaps that in some families all members do have the same shaped noses). The point in the case of games is that there obviously is no essence. What did Wittgenstein mean by 'family resemblances'? The example he gave was that of games [In the case of games resemblances are more sharp than blurry, however]: tennis and badminton both use nets (but not balls) and tennis and hopscotch both use balls (but not nets), and there are many more examples of these resemblances among games. And we may find this in a human family (although the shape of noses is rather vague [or blurry] compared with the presence or absence of nets): members of a family often look remarkably alike: e.g. there is the well-known photograph of young Wittgenstein with his brothers and sisters at table in Vienna showing a very strong facial resemblance among them all (and yet no two are identical). Or the members of a family may look much less alike: there are families where the children don't look much like each other, or where two may resemble each other, but a third not. And there are also families where an adopted child resembles a natural-born child although they have no common ancestor. -- But with respect to Wittgenstein's metaphor: the words 'adopted' and 'natural-born' are without application (i.e. that was not the comparison he was making). Which shows clearly that the expression "family resemblances" is a metaphor, not a theory about the cause of different things having been placed in the same category, as a theory about common ancestry would be; and that would have to be a metaphysical theory about an imperceptible essence in the case of games; because that games do not have an essence is a matter of fact (It is not a theory).

'Family resemblances' or 'family likenesses' need not be a single feature that is common to all cases: A and B may have the same shaped eyes, but different colored hair, whereas A and C have the same colored hair, but different colored eyes, etc., and in all these cases there may be greater or lesser resemblance -- or none at all! It is not obvious that board games like checkers [draughts] and field games like hockey belong in the same family (although, of course, anything can be compared to anything else in some specified way or another).

A classification scheme (which is what a common name is) may have a justification [in some cases there may be a common ancestry or a single common feature, as is the case with some categories in the sciences], but it also may not have a justification: there is not necessarily a red or scarlet thread running through, connecting all members of a class. We have inherited a natural language, and things are grouped together by it -- in whatever way we find them to be. It is not the aim of Wittgenstein's logic of language to invent theoretical justifications for these groupings (e.g. "abstraction of essences" or "recollection of Forms"): they are, rather, logic's given, "there -- like our life" (OC § 559). His logic of language describes whatever facts we find when we look -- and stops there. The demand for essences that philosophers make is not the result of an investigation but a presupposition they impose on their investigation of our concepts.

Does the family resemblances metaphor make anything clearer? It does if we feel the need for some way to characterize classes -- concept-words, class-words or common-names -- whose members do not all have something in common -- but only if we make its application clear by thinking it through for ourselves. Otherwise, this metaphor, like any other, will not be useful to us. Indeed, I think that it is clearer what Wittgenstein did not mean by this metaphor than what he did mean by it. But if we give examples and if we imagine drawing lines to join the common features, it becomes clearer what Wittgenstein had in mind.

Wittgenstein and Spengler

Wittgenstein "implies he got [the notion of "family resemblance"] from Spengler" (Wittgenstein in Cambridge: Letters and Documents 1911-1951 (2008), p. 9n16), although McGuinness does not say where Wittgenstein "implies" that. Maybe this remark is alluded to:

Spengler could be better understood if he said: I am comparing different periods of culture with the lives of families; within the family there is a family resemblance, while you will also find a resemblance between members of different families; family resemblance differs from the other sorts of resemblance in such & such ways etc.

But Wittgenstein does not say how to distinguish one family from a different family, nor how family resemblance differs from other kinds of resemblance.

... we have to be told the object of comparison, the object from which this approach is derived [because otherwise we shall] ascribe what is true [or, holds] of the prototype of the approach [or, of the comparison] to the object to which we are applying the approach as well ...

... the prototype must just be presented for what it is; as characterizing the whole approach and determining its form ... not by virtue of a claim that everything which is true only of [the prototype] holds to all objects to which the approach is applied. (CV (1998 rev. ed.) [MS 111 119: 19.8.1931]; cf. CV p. 14)

Query: Wittgenstein, rope without a red thread.

Then there is no way say whether there is a single rope ("family") or not (The allusion is to the defining red thread found running through all rope belonging to the British Navy [woven into the rope to discourage theft]), only to say that sections of rope look similar -- and therefore also dissimilar -- in various ways. Socrates' standard of judgment (Euthyphro 6d-7d) is absent (!). And you cannot skate over that absence on an empty metaphor.

"Fibers of a rope" metaphor

We find that what connects all the cases of comparing [i.e. the use we make of the word 'comparing' (p. 86)] is a vast number of overlapping similarities, and as soon as we see this, we feel no longer compelled to say that there must be some one feature common to them all. What ties the ship to the wharf is a rope, and the rope consists of fibres, but it does not get its strength from any fibre which runs through it from one end to the other, but from the fact that there is a vast number of fibres overlapping. (Wittgenstein, The Brown Book p. 87)

Or does its strength come from the force of habit, from the language we acquired from childhood, by which we feel that all the things we normally classify as games naturally belong together in the category 'game'? For a different upbringing, a different custom for using language from our own, might result in our regarding the placing of both volleyball and whist in the same category as eccentric and foreign to our way of thinking. And a different life-form from our own might be unable to conceive the fibers as joining together to form a single unit, and thus not have our concept 'rope' at all. Nor our concept 'game' -- Why, it might ask us in puzzlement, do you think that these particular resemblances (or, similarities) naturally group these different phenomena into a single category? It might say here: 'games' is a very strange category, for the differences between what you (i.e. our life-form) categorize as games are far more striking than the similarities.

In any case, "family" is a characterization after the fact, the fact being that our concept 'game' exists (It is part of our natural history), and if someone is so inclined (but then also so-disinclined (PI § 258)) he may characterize the similarities among the various games as "family" resemblances or "family similarities" or in any other way that strikes his fancy. To actually make a distinction (which is not a mere distinction without a difference), however, grammatical investigations must contrast common-nature concepts (words that are defined by a held-in-common feature) with non-common-nature concepts (words for which there is no general rule or standard for how to apply, but only a more or less complete list of examples (It is this list which a child learns when it acquires its native language and which, after that fact, it may feel to be natural rather than mere convention; feeling that way, of course, does not make it so)).

Do the remarks in Culture and Value "imply" that Wittgenstein got the notion of family resemblances from Spengler? I don't know. What is an example of a prototype -- i.e. an object of comparison -- Wittgenstein employs? He compares using language to playing a game, a game played according to fixed rules (e.g. a language-game we might describe or invent). On the other hand, would he say that e.g. chess is a prototype of 'game' despite its little resembling hopscotch? Are the differences from Spengler more or less important than any likeness to Spengler?

It does not, in any case, matter to philosophy if Wittgenstein got this notion from Spengler or from Mother Hubbard's cupboard. -- The question is: what did he himself do with it? See the Introduction: "... Wittgenstein, if he did not invent all the elements of his logic of language, so redefined them that they can only be gotten from him."

Just as, at this point in my study of philosophy, I am no longer concerned about whether I myself am using Wittgenstein's tools, that is, whether it is more or less, the way he did.

Spengler -- Wittgenstein was much impressed by Oswald Spengler's ... Decline of the West [1918-1922] ..., mentioning Spengler, alongside Sraffa, as one of the inspirers of his later philosophy.

Ever cryptic, McGuinness does not say where Wittgenstein "mentions" this. Maybe this remark is alluded to:

I think I have never invented a line of thinking but that it was always provided for me by someone else [and I took it up] for my work of clarification. (CV (1998 rev. ed.) [MS 154 15v: 1931 § 2]; cf. CV p. 19)

Wittgenstein follows that remark with a list of names of thinkers whom he says "had influenced" him [McGuinness uses the word 'inspiration' rather than 'influence']: Frege, Russell, Spengler and Sraffa (as well as Boltzmann, Hertz, Schopenhauer, Kraus, Loos, and Weininger; but this note is from 1931, and except for Spengler and Sraffa, these influences came from before the TLP; he also does not include Frank Ramsey in this list of "influencers" [His criticism was certainly "inspiration" to Wittgenstein], although Ramsey is named in the Philosophical Investigations).

(The idea of "family resemblances" is only the most obvious [But in what way, if any, is it "obvious"?] of Spengler's contributions.) To Drury [with whom he was agreeing in the first sentence McGuinness quotes below] Wittgenstein said, "You can't put history into moulds. But Spengler does point out certain very interesting comparisons." (Recollections p. 113) (Wittgenstein in Cambridge, editor's comment on p. 301 which is, McGuinness says, unrelated to the content of Letter 249)

But if Spengler does "point out certain very interesting comparisons", for what reason [or, purpose] are they interesting -- and does that have any necessary relation to defining words in logic of language? A comparison to Spengler's work is merely a comparison; an analogy is only proof of the possibility of making a comparison -- in some way or another (and neither Wittgenstein nor McGuinness says in which way).

"... into moulds", nor is language [always] a game played according to strict rules. That is a comparison one may make between Spengler and Wittgenstein, but what if anything follows from it?

• An exhibition at Cambridge University (27 April - 15 August 2011), Wittgenstein and Photography and Double Exposure Exhibition, has a different idea about where Wittgenstein got his metaphor 'family resemblances' from.

"What the sign slurs over, the use makes clear"

For many years I mistakenly believed this expression to have been invented by G.L. Farre and therefore I was afraid to use it, lest I steal someone else's work. Professor Farre, Georgetown University (fl. late 1970s), had, at least at the time I attended his lectures, what seemed to me a very strange idea: that Wittgenstein had never repudiated the "picture theory" of the Tractatus Logico-Philosophicus: the TLP is about the logic of propositions, whereas the Philosophical Investigations is about other forms of language.

I am now very grateful that my former teacher spoke mostly about the Tractatus, which was clearly his preferred text, -- because that forced me to come to an understanding of the Philosophical Investigations on my own. And that is always the best way in philosophy: to go by yourself to the original source.

What signs slur over, their application says clearly. (TLP 3.262, tr. Pears & McGuinness) What the signs conceal, their application declares. (tr. Ogden)

The word 'blur' -- 'slur' suggests spoken, whereas 'blur' suggests written language -- might also be used: "What signs blur over, their use makes clear." Ogden used the word 'blurred' in his translation of 4.112.

Site copyright © September 1998. Send Internet mail to Robert [Wesley] Angelo. Last revised: 19 May 2013 : 2013-05-19 (Original revision 27 May 2008)

The URL of this Web page:
https://www.roangelo.net/logwitt/family-likeness.html

Back to top of page

Wittgenstein's Logic of Language - Introduction and Table of Contents | Bibliography Site Search | Site Map