Wittgenstein: Family Resemblance

“Consider, for example, the activities that we call “games”. I mean board-games, card-games, ball-games, athletic games, and so on. What is common to them all?”

(Wittgenstein, 2009: §66)

When we categorise different things under a general concept – for example, categorising ‘football’ and ‘chess’ under the concept ‘game’ – we might imagine we are correct in doing so because the different things possess the essential characteristics of the general concept. Investigation into these essential characteristics – or essences – has been a feature of many philosophical questions throughout history such as:

  • What is knowledge?
  • What is art?
  • What is justice?
  • What is the good?

Typically, philosophical answers to these questions have attempted to provide a definition (e.g. necessary and sufficient conditions) or some kind of abstract object (e.g. Plato’s Forms) to explain what these general concepts mean.

But Wittgenstein casts doubt on the possibility of such an idealised conception of language in section 66 of Philosophical Investigations. Using the example of the word ‘game’, Wittgenstein hints at a different way of thinking about language that in later sections is described as ‘family resemblance’:

“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, colour of eyes, gait, temperament, etc. etc. overlap and criss-cross in the same way. – And I shall say: ‘games’ form a family.”

(Wittgenstein, 2009: §67)

The purpose of this essay is to contrast Wittgenstein’s idea of family resemblance with the idealised, so-called ‘Augustinian’, conception of language typically assumed in philosophy. Using a paradigm example from ancient philosophy – the definition of knowledge – I will show the philosophical confusion that results from assuming the Augustinian conception of language and show how this confusion can be avoided by replacing definitions/essences/objects with family resemblances.

The Augustinian picture

Wittgenstein attacks the idea that “the words in language name objects” (Wittgenstein, 2009: §1), which is how St. Augustine describes learning a language in The Confessions.

Augustine’s account of language may seem plausible when applied to ordinary physical objects – ‘table’, ‘chair’, ‘bread’ etc. – but runs into difficulty in accounting for general terms such as ‘game’. So, in order to maintain the Augustinian assumption for general terms, we must posit different kinds of objects to maintain this single uniform use of language.

One example of an attempt to maintain the Augustinian conception of language is to say that all games have the property of being a game, and further, that this property must be some kind of object:

“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. It is comparable to the idea that properties are ingredients of the things which have the properties; e.g. that beauty is an ingredient of all beautiful things as alcohol is of beer and wine, and that we therefore could have pure beauty, unadulterated by anything that is beautiful.”

(Wittgenstein, 1969: p17)

The Augustinian picture of language underlies a great many philosophical theories throughout history, perhaps most famous of which is Plato’s Forms. Plato posits an entire realm furnished with ideal objects – ‘Forms’ – to provide a (Augustinian) reference for general words:

“And we go on to speak of beauty-in-itself, and goodness-in-itself, and so on for all the sets of particular things which we have regarded as many; and we proceed to posit by contrast a single form, which is unique, in each case, and call it “what really is” each thing […] And we say that the particulars are objects of sight but not of intelligence, while the forms are the objects of intelligence”

(Plato, 2007: p232 – emphasis added)

In addition to Plato’s Forms, other philosophical ideas and theories that assume an Augustinian picture of language include:

  • Locke’s abstract general ideas1
  • Necessary and sufficient conditions/definitions
  • Essences

Even Wittgenstein himself falls under the spell of the Augustinian conception of language in his earlier work, Tractatus Logico-Philosophicus.

All these inquiries share the assumption that general words, such as ‘beauty’ or ‘knowledge’ or ‘game’, refer to some object – even if an abstract one – and that this object provides sharp boundaries for correct or incorrect use of the word.

The definition of ‘knowledge’

The tripartite definition of knowledge as ‘justified true belief’ at first appears to be one of the few philosophically successful examples of illuminating the meaning of a word through (Augustinian) definition.

Originating in Plato’s Theaetetus2, the conditions ‘justified’, ‘true’ and ‘belief’ appear to each be necessary, because a purported instance of knowledge would fail to be knowledge if it lacked one or more of these elements. Further, the conditions are jointly sufficient because there appear to be no instances excluded by the definition and the definition only includes instances of knowledge. The tripartite definition appears, then, to shed light on the ‘essence’ of knowledge independently of any particular instance of knowledge, and this definition has been hugely influential in the history of epistemology.

However, there are examples of ‘justified true belief’ that we would not typically consider to be knowledge, and these were first described by Edmund Gettier in 1963. An example of a Gettier case would be a man watching television whose phone, unbeknownst to him, is on silent. A phone rings on the television and this causes him to believe that his phone is ringing. However, by sheer coincidence, his own phone really is ringing (even though it is on silent). The man’s belief that “my phone is ringing” is justified (by the TV) and true, but is not knowledge – it’s merely lucky. Thus, such ‘Gettier cases’ show that the tripartite definition of knowledge is not, in fact, a sufficient definition.

Since Gettier’s counterexample, several philosophers have tried to salvage the tripartite definition of knowledge by modifying it3. But these attempts invariably share the assumption that there really is an essence that can be captured by definition.

Family resemblance

Wittgenstein rejects the Augustinian picture of general terms. Instead of assuming language functions solely to name objects, Wittgenstein argues that there need not be any essential feature to give meaning to general terms. His example of ‘game’ serves to illustrate this:

“Look for example at board-games, with their multifarious relationships. Now pass to card-games; here you find many correspondences with the first group, but many common features drop out, and others appear. When we pass next to ball-games, much that is common is retained, but much is lost. – Are they all ‘amusing’? Compare chess with noughts and crosses. Or is there always winning and losing, or competition between players? Think of patience. In ball games there is winning and losing; but when a child throws his ball at the wall and catches it again, this feature has disappeared. Look at the parts played by skill and luck; and at the difference between skill in chess and skill in tennis. Think now of games like ring-a-ring-a-roses; here is the element of amusement, but how many other characteristic features have disappeared!”

(Wittgenstein, 2009: §66)

The diversity of all the activities we deem ‘games’ shows the difficulty, if not impossibility, of providing a common property for all of them. From this Wittgenstein concludes: “we see a complicated network of similarities overlapping and criss-crossing: similarities in the large and in the small.” (Wittgenstein, 2009: §66)

The ‘family’ that constitute games may share various features between them, but need not all share any one feature, like in the following sets:

{A,B,C} {B,C,D} {C,D,E} {D,E,F}

We see here that ‘C’ is common to the first three sets but not the fourth just as balls may be common to rugby, golf and tennis but not chess. However, golf, tennis and chess share the feature of being non-contact whereas rugby does not. We would call all of these activities games, however, even if they are not united by any singular property.

Returning to the philosophical question “what is knowledge?”, we might likewise notice common features among instances of knowledge, but in other instances of knowledge this feature may be missing.

Don’t think, look!

Philosophers may be tempted to say that all games “must have something in common, or they would not be called ‘games’” (Wittgenstein, 2009: §66) but this is a philosophical temptation only. When doing philosophy, it is tempting to begin with the word and assume that its meaning is some abstract object.

But rather than theorising in this way, Wittgenstein urges “don’t think, but look!” (Wittgenstein, 2009: §66) because when we look at how a word is actually used in ordinary language, we see that no single feature is common.

Even if philosophical theorising were successful in future – say a definition of ‘knowledge’ was formulated that gave perfect necessary and sufficient conditions – it wouldn’t make a difference to how the word is used. In reality, we don’t really need to provide sharp boundaries for a word in order for it to be useful. This is demonstrated by the fact that, in non-philosophical contexts, there is rarely confusion about whether words like ‘game’ or ‘knowledge’ have been correctly applied:

“Can you say where the boundaries are? No. You can draw some, for there aren’t any drawn yet. (But this never bothered you before when you used the word “game”.)”

(Wittgenstein, 2009: §68)

The fact that we were “never bothered” that we couldn’t give the boundaries of ‘game’ before, and at the same time would say we understood the word, shows that the meaning of ‘game’ must be something other than necessary and sufficient conditions or some ideal object, as is typically assumed in philosophy. It’s the same story with ‘justice’, ‘beauty’, and ‘knowledge’: A definition isn’t needed for the word to be useful or for people to know what it means.

No more definitions

In summary, the way “similarities crop up and disappear” (Wittgenstein, 2009: §66) among different games suggests that there is no single definition or essence that constitutes the meaning of ‘game’. Further, the example of ‘game’ is analogous to many general words, for example ‘beauty’ or ‘justice’ or ‘knowledge’.

When Socrates asks “what is knowledge?” in Theaetetus, for example, Socrates expects as a response an all-encompassing answer (an essence, necessary and sufficient conditions, a definition, an abstract object, etc.) in isolation of any specific instances. So, when Theaetetus’ responds with examples, this answer does not satisfy Socrates:

“I asked for one [definition of knowledge], and you are offering many [examples]; I asked for something simple, and you respond with complexity.”

(Plato, 2004: p21)

However, Socrates’ demand for simplicity is unrealistic. Socrates is assuming an overly simplified, Augustinian, and false picture of how language works. General terms like ‘knowledge’ often don’t have rigid definitions but instead constitute a complex network of interrelated examples. Further, the fact that someone can’t provide a perfect definition for a word does not mean he doesn’t understand what the word means. Through examples, as Theaetetus gives, we come to understand how a word is used and thus what it means. The fact that in ordinary language we use general words, such as ‘game’ or ‘knowledge’, correctly, even though we can’t always provide sharp boundaries, shows that the meaning of such words is not some ideal object as the Augustinian conception of language assumes.

“What still counts as [knowledge] and what no longer does? Can you say where the boundaries are? No. You can draw some, for there aren’t any drawn yet. (But this never bothered you before when you used the word [‘knowledge’].)”

(Wittgenstein, 2009: §68)


1 E.g. “Ideas thus made up of several simple ones I call complex. Examples are the ideas of beauty, gratitude, a man, an army, the universe. These are all complex ideas made up of simple ones, but the mind can if it wishes treat each of them by itself as one unified thing, signified by one name.” (Locke, 1997: Book 2 Chapter 12 – emphasis added)

2 Socrates actually offers “knowledge is true belief accompanied by a rational account” (Plato, 2004: p117), which he later rejects. However, this seems to be the origin of the ‘justified true belief’ definition of knowledge.

3 E.g. Nozick’s reliabilist epistemology in Nozick, 1981.


Gettier, Edmund. (2000) Is Justified True Belief Knowledge? In Bernecker, S. Knowledge: Readings in Contemporary Epistemology (pp. 13-15). Oxford University Press.

Locke, John. (1997) Essay Concerning Human Understanding. Penguin Group.

Nozick, Robert. (1981) Philosophical Explanations. Harvard University Press.

Plato (2007) The Republic translated by Desmond Lee. Penguin Group.

Plato (2004) Theaetetus translated by Robin A.H. Waterfield. Penguin Group.

Wittgenstein, Ludwig (1969) The Blue and Brown Books: Preliminary Studies for the ‘Philosophical Investigations’. Basil Blackwell.

Wittgenstein, Ludwig (2009) Philosophical Investigations translated by G.E.M. Anscombe, P.M.S. Hacker and Joachim Schulte. Blackwell Publishing Ltd.

Wittgenstein, Ludwig (1961) Tractatus Logico-Philosophicus translated by D.F. Pears and B.F. McGuinness. Routledge and Kegan Paul Ltd.

Leave a Reply