Haig Khatchadourian—Common Names and "Family Resemblances"

Haig Khatchadourian
Common Names and "Family Resemblances"
Philosophy and Phenomenological Research, Vol. 18, No. 3 (Mar., 1958)

[SK's comments]

Maurice Mandelbaum:
Haig Khatchadourian has shown that Wittgenstein is less explicit than he should have been with respect to the levels of determinateness at which these resemblances are significant for our use of common names.

[341]

COMMON NAMES AND "FAMILY RESEMBLANCES"

I

... we propose to give, first,
a brief analysis of Wittgenstein's notion ...

Next we shall try to show
that

whether or not "family resemblances" constitute a general feature of ordinary language so far as common names are concerned,

there are at least some common names
such that the things named by them

do have

one or more features in common,

though this feature or these features are not

a determinate or relatively determinate

quality or characteristic.

Whatsoever could be a

feature

without also being

a determinate

quality or characteristic

?

Whatsoever
could this "determinacy"
be

relative

to?

Further,
we shall argue that this common "feature" or these common "features,"
in each case,
is or are what determine (at least in part) the applicability or uses of the common name concerned,

as is shown by an examination of the way in which the latter is used in ordinary discourse.

We shall next show how "family resemblances" themselves, where they obtain in the kinds of thing we are concerned with, can be accounted for in terms of our results.

Finally, we shall apply our results
to the terms 'good' and 'poor' or 'not good'
in one sense of these terms,
as used in ordinary language.

...

[342]

...

It seems clear that the "features" which Wittgenstein has in mind in speaking of "family resemblances"
are

determinate or relatively determinate characteristics

(or relations, or both),
and not merely

kinds of characteristic or determinables

(or kinds of relation, or both).

Is this to suggest something like basic level as against superordinate categories? That would be an impressive insight for 1958. But the alleged "determinacy" of one level "relative" to the other seems spurious.

Rosch says of superordinate categories that their "membership consists of a finite number of names of basic level categories which can be adequately sampled". So, it seems possible to argue that a "kind of characteristic" is not less "determinate" than any of its instances.

Failing that, there is the more basic question of what W is 'really' saying with the family resemblance thesis, and this seems to me to render K's attempted distinction moot. When we look for common properties with an eye toward establishing a 'necessary and sufficient' basis for category membership, it matters not whether we look for (nor whether we find) either "determinate characteristics" or "kinds of characteristics". In performing this task, we just are "determining", hence there really is no basis for attaching degrees of determinacy to our findings, "relative" to each other.

For we can imagine a "family" in which all the members have
a determinable or a set of determinables
but not
any determinate or relatively determinate characteristics,
in common.

All the members of a human family have eyes, ears, a nose, limbs, etc. in common;

but we would not (nor would Wittgenstein) say that these members have certain "family resemblances" by virtue of possessing eyes, ears, a nose, etc.

For the determinables these members have in common are shared by most if not all human beings.

Again, I assume that the "resemblances" which are operative in concept formation are 'gestalts'.

This is to say:
'gestalts' are neither

determinables

nor

determinate characteristics

per se.
All of that comes later. None of it is operative.

Of course it's true that

most if not all human beings

have eyes, etc., and this does point to a certain relativity in the very notion of resemblance, a relativity which we can assume has some part to play. What part, exactly? This just is family resemblance at a higher level of generality: we classify a legless man as a man, not as fish, tree, or sculpture; we probably don't classify him as tall, runner, or adonis.

Rather,
what Wittgenstein seems to have in mind is a "family" in which some members have
either

(1) one or more determinate characteristics in common
(say hazel eyes),
or
(2) relatively determinate characteristics in common
(say brown eyes of different shades),

and each of these members has one or more different determinate, or relatively determinate characteristics in common with some (at least one) or all of the other members.

So that some members are directly related by qualitative resemblances to other members,

while some or all are also indirectly related to other members through their direct relations to members themselves directly related to the latter.

But

how determinate

should a characteristic
shared by all members of a "family"
be

in order that it may be said to be "something in common" to all the members?

For

determinateness is a matter of degrees,

and is relative:

what is a determinate characteristic relatively to a given characteristic may be a determinable relatively to another characteristic.

Part of the answer here, if not the entire answer, seems to be that over- and under-inclusiveness must always both be considered. A 'hook nose' and 'dark curly hair' are more "determinate" than 'has a (any) nose' and 'has (any) hair', but these too are shared by huge numbers of nonmembers; these too are vacuous.

In other words, finding

"something in common"

may be quite easy; this 'something' may be highly "determinate" or highly "determinable"; and still, it may be vacuous vis-a-vis family resemblance if it is shared by significant numbers of nonmembers, as most such properties indeed are.

Perhaps W is not as explicit about this aspect as he could or should be; or perhaps it is all there in the Kampfspiele example. I will have to revisit this. It certainly seems that the PI sections that everyone quotes from are entirely concerned with 'under-inclusiveness', and this could indeed be a serious defect.

If Wittgenstein is merely repudiating the view that all things called by the same name (in the same "sense") have a qualitatively identical "determinate" characteristic
(or a set of these),
then

traditional essentialism will not be

completely overthrown;

since
on the latter view

relatively deter-

[343]

minate characteristics

which are

not qualitatively identical

are

counted as the "same."

It is obvious that it makes all the difference where we draw the line between

qualitative "sameness"

and mere

qualitative "resemblance."

If

anything short of

qualitative identity is rejected as a common quality,
even two copies of the same book,
or two dimes,
may only have

certain "resemblances"

but no quality

"in common."

On the other hand,
if the distinction is

drawn higher up,

so to speak,
in the determinate-determinable scale,
then the members of the family,
all of whom have brown eyes of different shades,
will have

a common quality

and not a "resemblance"

in color of eyes.

That Wittgenstein himself does not intend to draw the distinction too finely or too stringently is seen by considering his analysis of "games."

There he seems to consider games which are amusing as having a common feature;

similarly with games which involve
competition,
or winning and losing;

and so on.

Yet "amusiveness," for instance, may be of different kinds.

Nevertheless, the difficulty seems to remain:

for

on what logical basis or bases

(whether the same or different ones in different cases)

would Wittgenstein

draw the line

in a given case,
and in different given cases,
between
qualitative "resemblance"
and
qualitative "identity"?

Well, perhaps he doesn't have to (doesn't think he has to).

i.e.

The

line

is

predrawn

by 'language'.

There is another point to note
before we proceed further.

This is that in talking about "family resemblances," Wittgenstein is concerned with things which are

called by the same name

in one and the same sense,

and not in different (literal) senses.

Obviously, if what Wittgenstein is maintaining is simply that things which are called by the same name. but in different senses, or with different meanings, have only "family resemblances" ... , his view would be of relatively little significance.

For hardly anybody would hold the contrary, ...

We shall now analyze the uses or meaning of the name 'games,'
in some detail,
as a paradigm for the view we shall argue for;

and then apply our results to one class of objects,
namely
manufactured — in general, manmade — objects.

Wittgenstein seems to me to be quite right in holding that there are no determinate or relatively determinate characteristics common to all things called "games."

All the games that we have, and all phenomena that we would normally call "games," are played in accordance with

[344]

certain kinds of rule;
so that there is a universal common to them all;
but the rules involved differ as we pass from board games to ball games to card games, etc., and from one boardgame to another, from one ball game to another; and so on.

Still,
there is

a more determinate feature common

to all kinds of game:
namely,

the capacity to serve a specific human need or needs,

directly or indirectly, under what we shall call "standard" (causal) conditions or in "normal" contexts.

This is over-inclusive.

The term 'need' is used here in a wide sense ...

Instead of speaking of
the "capacity to serve a human need," we might speak of
the "capacity to further human ends,"
or
the "capacity to produce certain effects
in other things, which, directly or indirectly, serve or can serve human ends."

Or we might speak of the phenomena or things whose "capacities" we are talking about as having or as capable of having a specific use or uses for human beings under specified conditions.

In the case of "games," this "capacity" is the capacity to evoke or produce pleasure in the player or players and/or the spectators.

That this is implicit in some way to be determined later in the meaning or uses of the name 'game' can be seen in a general way by considering some remarks we ordinarily make in connection with "games";

e.g.

"What sort of game is this? — it's boring!";
"He always gets angry while playing — and spoils the game";
"This is certainly a queer kind of game:
I don't see what pleasure you (or the players, or the spectators) get out of it!";
"You call this a game?
Well, you may call it that if you like,
but it's certainly a poor game:
it's positively irritating!"

. . . 'You call this art?' . . .

Now here is a

remark

we ordinarily make.

I don't agree that we "ordinarily" say this about games, nor about spouses, ankles, or forks; yet we can easily imagine situations where this would be said of these things.

If there is any generalization to be made here about what we say, it is indeed a function of how we live, and it is not a function of what all of these things are in some more ultimate sense.

Also, we do not normally say: "This is a pleasant game" (though we quite frequently say: "This is a very pleasant game"), except perhaps to mean that I liked it, or that I find it pleasant.

"This is a most unpleasant game" is possible though unusual, and makes one think of games which some would find coarse or vulgar ...

In the last case we tend to think that the term 'game' applies only in a loose or extended or peculiar sense, or that the person making the statement is prudish, that he is judging the game in terms of extrinsic criteria, ...

"I found this a most unpleasant game"
would, however,
not be puzzling.

But "This is a painful game," seems to be almost a contradiction in terms

in "standard" contexts.

It raises questions in the hearer's mind, and

requires explanation.

(It will be explained, however, if the game referred to is bear-baiting, say, and the speaker is thinking of the bear's suffering, or of the pain he feels for the bear's suffering, or of

[345]

both.)

A new version of the same mistake.

If this is the situation,
then no

explanation

is

required

.

Certainly we can and often do fail to understand each other in precisely this way, but establishing the possibility of misunderstanding does not establish the possibility of its remedy. If I fail to 'just know' what you mean in calling a game "painful", I fail absolutely: no "explanation" can deliver me to the state of 'just knowing'; the best I can now do is to settle for 'intellectual understanding' in lieu of 'just knowing'. This is often necessary, but it is an imperfect substitution.

If someone says:
"This game made me angry,"
we think:
"He (or his favorite team, or player) lost, that's why";
or that he did not play as well as he knew he could, so he felt frustrated.

If a person does not enjoy a game,
and we

regard the conditions

as "standard" conditions so far as the majority of spectators and/or players are concerned,
we will think:
He's not "in" it;
he hasn't caught the "spirit" of the game;
he doesn't have the "right" spirit;
...
and so on.

We don't really

regard the conditions

until after we already

think

what we're going to think.

In order to understand these remarks
in the light of what we said about "games,"

we have to consider our crucial qualification that what we call 'games" have a common capacity to produce pleasure under standard conditions or in normal contexts.

What we ordinarily call "games" certainly do not actually produce pleasure under just any and all conditions, in all contexts.

It is only when certain conditions obtain that a "game" actually produces pleasures.

This is why we said that "games" have the "capacity" to produce pleasure under "standard" conditions.

The word "capacity" is meant to indicate that if certain conditions obtain, pleasure will⁴ be produced.

---

⁴ The statement: "If standard conditions obtain, pleasure will be produced by a game" is meant to be, and is, an analytic statement; and not either an a priori synthetic judgment, or an empirical generalization.

For our contention is that

the notion of "standard" conditions

here is

built into the meaning

of the word 'game' in such a way as to make the above statement analytic.

The essential point here is that we are not

arbitrarily introducing the notion

of "standard" conditions in order to

guarantee the truth of the statement.

Our contention is that if we analyze the uses of the term 'game,' we discover this notion of "standard" conditions implicit in it.

That this is so is, however,

an empirical assertion

about the uses or meaning of 'game.'

...

---

This seems obviously wrong, and it marks a familiar limit of

empiricism.

Namely: in such arenas as

meaning

we can never be sure that we've empirically accounted for any and all so-called

non-"standard"

conditions.

Why?

Because, as far as we can tell, the scope for such things is infinite.

Then, if

the notion of "standard" conditions

was

built into the meaning

of a word, that word would not travel nearly as well as most words do.

Perhaps it would be better to speak of the 'fit' between "words" and "conditions"?

To put the matter without using the misleading word 'capacity,'

what we are saying is that
"games,"
under certain conditions,
produce
(either physically or psychologically, or both:
in general, causally)
by virtue of whatever characteristics they possess,
what we call "pleasure" in the player or players, etc.

The word 'capacity' is not meant to stand for a mysterious power or property, over and above the ordinary characteristics of a game.

If we were not unwilling to complicate matters by using controversial notions, we would say that games have a certain "dispositional property," or a set of these — the property or properties of producing pleasure.

In making the statement

"This is a good game,

but I don't enjoy it,"

the speaker tacitly recognizes or implies
that
although he is not being affected in the way which is normal in those cases in which he would, without hesitation, call a given phenomenon a "good game,"
the absence of this element in this case is not a sufficient ground for not calling the

[346]

phenomenon concerned a "good game."

In saying that it is a game or a good game,
he concedes that it can give,
and perhaps is actually giving, others,
some, or even a good deal of, pleasure;

and also,
he implies that his failure to have pleasure is not due to the "game" itself, but to his mood, state of mind, etc. at the time.

Thus if somebody asked our friend:
"Why do you say so?",
on hearing him make the above statement,
the answer generally given will be something like:
"O, I'm not in the mood to-day";
or,
"I have a toothache";
"I'm worried about my affairs";
and the like.

That is, the speaker implies that so far as he is concerned, the conditions are not normal or standard conditions.

Well,
now we have over-fit the

conditions

to the

statement.

i.e. This sort of "statement" is made, but it may be made for reasons quite other than the ones given here.

But now a doubt arises:
is the pleasure produced by different games,
or by different kinds of game,
of the same kind in every case?

Are not the pleasures produced related merely by certain "family resemblances," at least in the case of the pleasure produced by games of different "kinds"?

If the answer is in the affirmative — and Wittgenstein would say that it is in the affirmative — then obviously Wittgenstein's analysis of games would be completely true,

though
here it would be

the effects of games

that are related by "family resemblances,"
rather than,
or as well as,

their characteristics

themselves.⁵

Now it is certainly true that different kinds of pleasure may be involved in the case of different kinds of game:

...

But though only some kinds of game produce or can produce physical pleasure, it seems that all games produce or can produce psychological pleasure — though the amount and intensity of pleasure produced may differ ...

But are all these psychological pleasures of the same kind in all cases?

No, again.

The nature of the specific game determines in good measure the kind of psychological pleasure derived.

...

... psychological pleasure may be divided into sub-classes, which are themselves further divisible.

Nonetheless, it seems to me
(though it is not possible to defend this here),
that the classification into "intellectual" and "non-intellectual" (emotional) pleasure
more
concerns the way in which the pleasure is produced and its general intensity
than
marks a

[347]

distinction in the kind of experience involved.

Whereas, a difference in kind of experience is involved in the case of the distinction between psychological and physical pleasure, in addition to a difference in the kind of way they are produced in us.

Thus

a relatively determinate kind of effect

seems to be produced

by all kinds of games.⁶

---

⁶ It is noteworthy that many of the phenomena Wittgenstein analyses are complex "psychological" processes, such as thinking, doubting, learning;

and it may well be that there only "family resemblances" can be discovered.

But supposing that to be true, and even if in the case of games themselves, only "family resemblances" are discoverable in their effects,

our position seems to remain secure as far as the names of manufactured objects are concerned — and also of activities which do not appreciably include psychological elements in their effects.

---

And so far as this relatively determinate kind of effect is produced by all games, it seems to follow that there is

some kind of "capacity"

in the games themselves,
which is
(at least)
similar in the case of all games;

though there need not be
(and we are not maintaining that there is)
any

determinate or relatively determinate characteristic

or characteristics common to all games, by virtue of which this kind of effect is produced.

What we have said applies,
mutatis mutandis,
to all other man-devised activities,
to man-devised processes,
and to man-made objects,
...

...

Here K expresses his argument in Alphabet Soup format. I choose to excise most of this.

... the notion of a common capacity Ca, or alternatively, of a common use U, is implicit in the meaning or uses of 'X' in ordinary discourse.

It is true that activities, processes, or objects called by different names may satisfy the same (kind of) need, may have the capacity of being used in the same (kind of) way;

also, that activities, processes, and objects which are called by the same common name can be

[348]

used in different ways.

We can use a table as well as a chair for sitting, and we can use a chair as well as table for eating or writing.

But sitting on tables is not

ordinarily

regarded as

the normal or proper way

of using tables,

nor eating off a chair or writing on it as the normal or proper way of using chairs.

(I am not using 'proper' in the sense in which it is used in

books on etiquette.)

File this

etiquette

with Professor Wisdom's "question of law, not of fact".

Perhaps
another way of saying
that
(a)
we can use a chair

for eating

though
(b) it is not

normal

to do so
is
to say that
there are situations in which this would be readily understood, and other situations in which it would be perplexing.

We can only assume that if certain situations occur often while others occur hardly at all (or not at all), convention will indeed reflect this. But again, our ability to apply the word in new situations, seemingly open-endedly, and seemingly with no loss of

propriety,

is a bit of a problem for any argument which generalizes from convention.

...

...
it is

we, the makers

of tables and chairs, who

design them in such a way

that it becomes possible for them to satisfy the need, or to have the use in view, more efficiently or better than another or other needs or uses.

This, as we said, is reflected in ordinary language:

the notion of some proper function or use or another (or of the capacity to satisfy a given need) is implicit in the meaning of the common names — of the particular sort we are talking about — which we give to things.

It is no accident that
the names of

manufactured objects

(as also the names of

man-devised activities,

or processes)

implicitly involve in their meaning

the notion of a capacity

to satisfy a specific need or specific needs,
to have a given use or uses.

"Separate seat for one,"
a common definition of 'chair,'
illustrates this clearly.

In many cases, however,

dictionaries

give a description of the characteristics which the things called by the given name were

endowed with

by their inventors or original makers
(hence, when the particular name

was first given

to them),
and/or those characteristics which they

most frequently

possess,
in addition to mentioning a specific use or specific uses.

...

Further,
as in the case of "games,"
it is seen that no single determinate or relatively determinate characteristic, or set of characteristics, common to all things ordinarily called "chairs,"
is discoverable.

The same is true of things called "tables," "pens," and so on.

An object

may be

made of chromium or of plastic, may be high or low, soft or hard, round or square or polygonal, straight or curved:

and yet

will not,

for that reason,

be refused

the name 'chair' — so long as it

can serve

as a "separate seat for one";

i.e., so long as the variation in these and other characteristics does not grossly impede or make impossible the object's possession of the requisite capacity.

If an object is made of jelly,
it

cannot serve as

a chair;
and we do not

apply 'chair' to it

in a straightforward

[349]

sense,
even if it

has the form

of one kind of chair or another.

This seems wrong, and art is indeed a place to start looking for the reason why.

What is necessarily not

straightforward

in calling a jellychair a "chair"?

We may imagine a contemporary art-piece that presents as, and is described as, and is spoken of as, a chair made of jelly. I stop short of saying that it is a chair, only to avoid suggesting anything ultimate; but I think it's totally wrong to say that this

application

of the term is not "straightforward".
Certainly this use is

infrequent,

but that is something else entirely.

Now, does this example show, actually, that art is the single worst place to look for the particular answer we're seeking? I'm not sure, but I think we can find art-ish examples everywhere, on a continuum of sorts from K's pragmatically-determined ones to my imaginatively-determined ones . . .

Similarly
chair-like objects only a few inches wide and high may be called "toy chairs," but not "chairs" without qualification.

They may be called "toy chairs" because
(1) they have the over-all form of one kind of chair or another,
and
(2) they can serve to seat
tiny "toy men" and "toy women" ... , or little

imaginary

creatures.

How much variation in form and material is compatible with the requisite capacity, depends on the nature and range of the corresponding "standard" (causal) conditions, and on the nature and number of the needs the thing is intended to satisfy.

In general,

the more specialized and the more complex

the need,

the smaller the range of variation

possible;

the less specialized and the less complex

the need,

the greater the range of variation

possible.

Similarly,
the range of variation is inversely proportional,
roughly speaking, to
the number of needs

meant to be

satisfied.

The form of chairs designed

merely with "utility" in view

admits of

considerably greater variation

than that of chairs designed with an eye to

beauty as well as to "utility";

and chairs designed only with an eye to beauty
(a mere hypothetical example, however!)
admit of greater variation in material than chairs designed with an eye to both beauty and "utility."

I'm very skeptical of the notion that these

"toy chairs"

simply would/could not be called

"chairs"

without qualification.

If we iterate "chair" examples until we have a suitably nuanced spectrum of possibilities, I think we find, first (if nothing else), that the Principle of Least Effort has a significant part to play; next (and more importantly for present purposes), we find the sway of the 'total situation':

e.g. perhaps when a child becomes suitably engrossed in toy-play, all "toy chairs" become simply "chairs";

when the child is somewhat older, perhaps the science-project mini-chair made of toothpicks, which is no good for sitting yet is quite explicitly

man-devised

and

designed,

is simply spoken of as a "chair" so long as the science project is front of mind in the given situation.

Now, if derelict Uncle Steve has been evicted and is crashing with the family while he gets back on his feet, perhaps he is the one who comes across the toothpick-chair and blurts out: "you call that a chair??" But this is easily explained.

The contrived nature of this example is precisely the point of it: this much distance from the 'situation' is required, I think, to bring K's theory to life.

What the 'iteration' of chair-examples shows about

capacity

is that capacity can be imagined, projected, remembered, etc.

Certainly it can also be

meant,

but such talk is cheap.

i.e. Both "imagination" and "intending" ("meaning" in the sense of "intending") are infinite in the Cartesian-Chomskyan sense, but only one of them has the final say.

e.g. I 'imagine' I am unable to do much hitting either with Babe Ruth's bat or with a souvenir mini-bat, but I do "apply" bat to both, so long as we are talking baseball, and I do 'imagine' that I could do some hitting (though not much) with either if I was simply forced to attempt this.

From all this we see why no determinate or relatively determinate characteristic or characteristics need run through all things called by the same name for that name to be applicable to them.

Hence it is not surprising,
if one fixes his attention on the determinate or relatively determinate characteristics of things,
that he tends to conclude that things called by the same name have no determinate "feature" or "features" in common;
or that, at most,
they have certain "family resemblances."

It is especially easy to fall into this line of thought since we become aware or fully aware that a common capacity is or may be involved only by considering the characteristics of a thing from a very general standpoint
(i.e., by looking at them as general kinds of characteristic or as determinables),
or by considering a thing in the light of the way we talk about it in actual contexts, under both "standard" and "non-standard" (causal) conditions.

Well, there seems to be a serious obstacle to thinking that merely

considering

this can lead to a conclusive answer.

Accounting for only these

actual contexts

doesn't get us very far at all.

For as we have said, things which do satisfy a common need, or have a common use, in certain kinds of situation, may not be capable of satisfying the need or of having that use in other kinds of situation.

II

The preceding —
and especially the fact that the possible range of variation in the characteristics of a thing is, roughly speaking, inversely proportional to the complexity and specialization, among other things, of the uses it is ordinarily put to, the ends it serves

— gives us one explanation of why we do find "family resemblances" between things called by the

[350]

same name — the feature on which Wittgenstein fixes his attention.

The clearest way to show this is by considering things like chairs, tables, cups saucers, hammers, nails, and the like.

In the majority of cases we unqualifiedly apply the name 'chair' to objects which have a certain kind of over-all structure and shape, and are made of kinds of material falling within some range.

This general, over-all pattern

imposed by the purpose

for which chairs are designed
— serving as a separate seat for one —
gives things generally called chairs

certain over-all similarities in form,

and to a lesser degree, in kind of material.

A "chair" is expected
to have some kind of seat,
to rest on some kind of support,
to have some kind of support for the sitter's back; ...

...

But different chairs may serve different, more specialized needs, ...

The kinds of specialized need,
or the kinds of specialized use,
give us the basis for the ordinary classification of chairs into kinds.

They impose further restrictions on the form and material of chairs:
restrictions over and above those imposed on them by the end-in-view of serving as separate seats for one.

These additional restrictions give rise to greater resemblances between

chairs of the same "kind"

(chairs which satisfy the same kind of specialized need, which have the same kind of specialized use),
and to a lesser extent,
between chairs of different "kinds"

whose specialized uses overlap

or are similar in certain respects.

Thus a "reclining chair" bears greater resemblances to other "reclining chairs" than to ordinary, plain chairs.

But "reclining chairs" may be [further] subdivided ...

These resemblances in detail between chairs of the same "kind," and more general resemblances between chairs of different "kinds," give us at least some of the overlapping and criss-crossing of resemblances which Wittgenstein speaks about.

...

The same applies to games,
as we see when we consider the general kinds of game:
...

But the general purpose meant to be served by a

[351]

game,
as a game,
and the more specialized purposes meant to be served by it as a specific kind of game,
determine its characteristics to a lesser degree than the corresponding purposes in the case of
a chair, say,
because of the former's

greater complexity

and

greater indeterminacy.

...

III

In talking about games,
we gave examples of the kinds of condition which are "standard,"
and of others which are not "standard,"
relatively to games in general.

It is now time to say something more general and more precise about these.

The "standard" conditions, S₁,
relative to a thing T,
and to a given use U, are:

(1) The kind of condition
— physical, or psychological; or both —
which are causally necessary for T's having U by virtue of its (T's) characteristics.

"Non-standards" conditions, S₁, relative to T and U, are the kinds of condition — physical, or psychological, or both — which causally impede, or even make impossible, T's having U;

assuming that T itself, as T (i.e., as called "T") is not what we would ordinarily call defective.

For even under "standard" or "favorable" conditions relative to T, a thing T, lacking some part it ordinarily possesses, may be incapable of having U.

Thus even under "standard" conditions, relative to chairs, a chair with a missing leg, for instance, cannot, as it stands, be used as a chair.

So, a soaked football field, a tilted stage, and a zero-gravity spacewalk would all be

non-standard

conditions

vis-a-vis chairs.

What about the skeuomorph-retronym phenomenon?

(2) There is no hard and fast dividing line between "standard" and "non-standard" conditions relative to T and U.

Also, there does not exist a unique set of conditions which is "standard," and another unique set of conditions which is "non-standard," relatively to T and U.

A good deal of flexibility is possible; though mainly in situations where psychological factors are important.

A pen cannot be used as a writing instrument without some kind of writing fluid;

but a game of baseball may be played and enjoyed even in relatively bad weather, if the players and spectators are in the mood for playing or watching, respectively.

(3) In the case of any given common name 'X'
(of the kind we are concerned with in this paper)
there are certain conditions,
fixed by linguistic usage,
which determine whether or not 'X' shall properly apply to a given thing T:

conditions under which we would refrain from calling a thing an "X" if it did not have a use U the notion of which is implicit in the meaning of 'X.'

These conditions may be called
the "standard conditions for the use or application of 'X' to a thing T" (symbolized by 'S₂').

In general, S₁ and S₂ coincide.

More clearly, conditions S₁ under which a

[352]

thing T normally has a use U are, generally speaking, the conditions under which T's having or not having use U becomes crucial for determining whether or not a name 'X'
(which involves the notion of use U in its meaning)
is applicable to it.

If T does not have U under standard causal conditions S₁ relative to X's in general, 'X' will not, generally speaking, be applicable to it.

But
S₂ and S₁
do not always coincide.

For instance, there are cases where we do not refrain from calling a given thing an "X," even though it is incapable, as it stands, of having a use U under the standard causal conditions S₁ relative to X's in general, and though the notion of U is implicit in the meaning of 'X.'

For example, a "broken chair," or a "chair with a leg missing," is called a chair even though it cannot be sat on under causal conditions which are standard for intact chairs.

The existence of such cases may seem to contradict our general position in sections I and II.

But actually it does not materially affect it.

As a matter of fact, such apparent exceptions

merely constitute borderline cases

and are themselves explainable in terms of our general position.

A. Let us take a simple and clear example,
that of chairs.

Chairs seem unmistakeably to have been invented for a specific kind of use, for use in a certain way.

That is,

the intended use

seems to have determined
the general characteristics of

the original chairs.

It is possible, therefore, that originally the name 'chair' did involve, in ordinary discourse, the notion of the capacity to have that use under certain physical conditions such that a thing could not be properly called a chair unless it possessed that capacity.

But since

a chair's characteristics

are causally responsible for

the chair's possessing the capacity

to be used in a given way,
under certain conditions,
the term 'chair' would subsequently come to apply to objects which have these kinds of characteristic, in their own right,
making

the possession of these kinds of characteristic

a "necessary" condition for the applicability of 'chair,'
rather than the capacity to have

the use which these characteristics

are designed to make possible.

Once this shift in "meaning" has been effected,
a further complication seems to arise.

The term 'chair' now comes to be applied also to objects which

merely "resemble"¹¹

objects hitherto called chairs,
provided the "resemblance" is considerable and not

[353]

tenuous
(there being no sharp line of demarcation between "considerable" and "tenuous" resemblance),
even though these objects,
as they stand,
cannot serve as chairs.

---

[352]
  ¹¹ I am not here referring to the kind of resemblance which obtains, say, between so-called "irrational numbers" and so-called "cardinal numbers,"

on the basis of which

[353]

(taking this merely as a hypothetical possibility)
the former may have come to be called "numbers."

This kind of extension of a name's meaning does not consist in the creation of borderline cases, as in the case of a so-called "broken chair's" being called a chair despite its being "broken."

Incidentally, the former is one way in which the existence of "family resemblances" between things called by the same name may be accounted for.

---

By an "object which 'resembles' a chair" I have here in mind a so-called broken, battered, or incomplete chair, such as a chair with one leg missing.

B. But the shift in "meaning" described above
is not sharp and complete in actual cases.

The notion of certain sets of characteristics Ca, or Cb, or Cd, and so on, does not seem to usurp completely, in actual practice, the place of the notion of a given use U which Ca, or Cb, or Cd, and so on, make possible in a given thing T.

In deciding whether or not to call a given thing by a certain name, or in justifying the giving of a certain name to a given thing, appeal is made now to the one, now to the other, now to both, notions.

When we say:
"What sort of chair is this? — it's broken; one of its legs is missing!"
our tendency to think of the object referred to as a broken chair may be due either to the fact that

(a) it lacks certain kinds of characteristics which are generally found in objects called "chairs" without qualification
(so-called intact chairs),
and/or to the fact that
(b) it cannot, as it stands, serve as a chair.

But we do speak of it as a "chair" because we think of it

(i) as an object which used to have the kinds of characteristic a chair generally has, or was capable of serving as a chair; or both;
and/or
(ii) as an object capable of serving as a chair if it were to have certain characteristics it now lacks (i.e., if it were to be "repaired").

Thus if (say) just before a party a servant discovers a chair with a cracked seat, and he remarks:
"We can't use this (as a chair): the seat is cracked,"
he may get the following reply from the hostess:
"Yes we can! Get me some glue,"
or
"Yes we can, if you get me some glue."¹²

---

¹² It is important to note that in the above interchange 'can' is used in two different senses by the servant and the hostess respectively.

The former is using the word in a narrower and stricter sense.

He is saying, in effect, that as it stands, and under the prevailing circumstances, the "chair" cannot serve as a chair.

The hostess, on the other hand, is using 'can' in a less stringent sense.

What she says, in effect, is that the broken chair can serve as a chair, even in that particular situation, since it can be (easily) repaired, and therefore made to serve as a chair.

In this sense of 'can,' our statement that an object (activity, process,of the kind we are concerned with) may be called "X" even if it cannot serve as an "X," etc., will be false.

In this sense of 'can,' "can serve as an X" can be taken as a (conventionally determined) "necessary" condition for properly calling a man-made or a man-devised thing an "X."

However, let us add that the ordinary uses of 'can,' in our second sense, are relative to the particular situation involved. A

[354]

chair with a slightly cracked seat "can (be made to) serve" as a chair, even in the urgency of preparations for a party;

but a chair with a leg missing cannot (even in the second sense of 'can').

So the hostess would not say:
"Yes we can, get me some glue!"
if the chair referred to had a leg missing.

(But she would still refer to the object as a "chari.")

Still, a chair with a leg missing "can (be made to) serve" as a chair if there is plenty of time, say, to send it to a carpenter and have it repaired.

Thus the distinction between 'can' and 'cannot' in this sense is not clearcut and sharp:

whether or not a given object is thought of as capable of having a given use depends on the state of the object in a particular situation.

---

In using qualifying adjectives like 'broken,' 'defective,' 'incomplete,' 'battered,' we reflect in ordinary language the distinction between the kind of case in which a given object

[354]

called "X" can, as it stands, serve as an "X," and the borderline kind of case in which it cannot, as it stands, serve as an "X."

Let us note that there is no hard and fast line dividing the borderline cases themselves, where the name 'chair,' say, is still applied (though with qualification), and cases where the name 'chair' is no longer applied.

Thus we uniformly speak of a "chair with a leg missing," a "chair with two legs missing," a "chair with three legs missings";

but we can either speak of a "chair with all of its legs missing," (or a "legless chair") or of the "seat and back of a chair";

while we do not seriously speak of a "chair with the legs and back missing" but rather of the "seat of a chair," or even of a "wooden board" (or a "wooden board having the shape of a chair's seat"), and the like.

In other words, and stated generally, there is no hard and fast line in ordinary discourse between an "X with a part missing," a "part a of X," and a "not-X and not-a."

IV

Our discussion would be grossly incomplete if we do not consider,
at least briefly,
one use of 'good'
and of 'poor' or 'not good';

namely,
their use in such statements as
"This is a good knife,"
"That wasn't a good game of tennis,"
"He played a poor game of chess."

Such an analysis shows that there is a very intimate relationship
between
the uses of common names of the kind we have concerned ourselves with in this paper,
and
the above uses of 'good' [etc.] ...

A. If we examine cases in which we ordinarily call a man-made object, or a man-devised activity or process, a "good X" and not merely "X,"
we find,
first,
that we do so

both

(a) when the prevalent causal conditions
(viz., in general, the conditions for the application of the name 'X')
are standard,
and
(b) when the causal conditions are non-standard, for "X's" in general.

In the case of (a) we further discover that in calling the given thing a good "X,"
we indicate one or both of two things:

(i) that this thing "X"
is capable of having, or actually has,
in an eminent or outstanding degree
(these terms being used in a relative sense)
a given use U;

(ii) that

[355]

"X" possesses one or more characteristics by virtue of which it is capable of having, or actually has, a use U, in an eminent or outstanding degree.

In both (i) and (ii), the use U involved is found to be the use associated with "X" qua "X" or the notion of which is implicit in 'X's' meaning.

For instance, if we point to a knife and say:
"This is a good knife,"
we ordinarily indicate that it is capable of cutting, or is actually cutting, certain kinds of thing (meat, butter, bread, and so on) well, efficiently, without effort,
and/or it is either sharp or very sharp, or has a firm and comfortable grip, or has a long and wide blade, or its blade has a saw-like edge, or some or all of these characteristics together — that is, one or more characteristics which make it capable of cutting certain kinds of thing well, efficiently, with ease.

We may also speak of a knife as a good knife if it has one or more characteristics that make it capable of being used to cut (of serving as a knife), or to cut well, for a relatively long time:
in general, characteristics which enable it to preserve its peculiar use.

Thus a knife may be said to be a good knife because it has a rust-proof blade and/or has a sturdy handle.

Examples illustrating these points
can be easily multiplied.

Also, these points can be readily applied, mutatis mutandis, to the use of 'very good, 'excellent,' and the like, ...

In the case of (b), i.e., when we apply the term 'good' to a thing "X" under causal conditions which are non-standard for "X's" in general, we indicate that 'X" is capable of having, or actually has, use U even though the prevalent conditions are non-standard for it qua "X," and therefore even though it is not expected to have U.

If "X" has U, or seems capable of having it (or if it has characteristics which seem to show that it has this capacity) in an outstanding degree, even though the prevalent causal conditions are non-standard, we tend to call it a "very good X," or an "excellent X," and not merely a "good X."

We will call a knife a good (or a very good) knife if it cuts (and better still, if it cuts well) things which it is

not designed

— and therefore not expected —

to cut.

Similarly, we would ordinarily apply the term 'good' (or 'very good,' etc.) to an all-purpose knife just invented, tacitly comparing its performance with the performance of ordinary knives.

But once we get used to this kind of knife and ordinary knives go out of use or become rare, we tend to bestow these terms only on a specific all-purpose knife (or a whole brand of it) which cuts better the things all-purpose knives are expected to cut, or cuts more kinds of things than other all-purpose knives, under conditions which now come to be regarded as standard for them, but which include conditions which are non-standard for ordinary knives.

...

This example also illustrates the converse of what we have just said:

we tend at present to speak of a quill pen as a poor pen, judging

[356]

its performance in terms of the performance of fountain pens, when originally it must have been the opposite.)

It may be noted that in case (b) in general,

(i) the degree in which a thing "X" has or seems to have use U associated with it qua "X,"
and
(ii) the degree in which the prevalent causal conditions are non-standard (or unfavorable), wherever such variation in degree in these conditions is possible,

ordinarily determines the judge's assessment of how eminently "X" has or is capable of having U.

Consequently, they determine whether 'good,' or 'very good,' or 'excellent,' shall be applied.

These points may be illustrated by the following statements:

"It was a very good game of tennis!
The players rarely missed the ball, despite the strong wind";

"It was an excellent game of tennis!
The players rarely
(or better still: "hardly ever," or "never")
missed the ball,
and did some fast footwork,
despite the strong wind and the soggy ground."

Our remarks in (a) and (b) above
apply equally, mutatis mutandis,
to objects, activities, and processes
with relatively, or highly, specialized uses,

such as carving knives, chess games, sledge hammers, writing desks, armchairs, ice cream, and so on.

B. Coming now to uses of 'poor' or 'not good' in the sense in which we are here concerned with these terms,

a man-made object or a man-devised activity, or process, is ordinarily said to a be a "poor X" or "not a good X" in the following situations:

(1) If it is regarded as completely incapable of having a given use U associated with it qua "X,"
provided that

(a) it is regarded as intact qua "X,"
and that
(b) the prevalent causal conditions are regarded as standard,
that is, when "X" is expected by the person calling it a "poor X" or "not a good X" to have use U.

(In this type of case, however, we generally tend to use a stronger expression than 'poor' or 'not good' alone:
we say:
"This isn't a good X at all,"
or
"This X is useless,"
or
"This is a very poor X.")

A knife with a completely blunt blade is intact as a knife — it has the general structural characteristics of one kind of knife or another, though not all the characteristics which are necessary for cutting.

Hence it is called a knife, and not, say, "junk" (except as a hyperbole).

But since it is incapable of cutting things which knives are ordinarily expected to cut, it is called a poor knife or is said to be not a good knife.

(On the other hand, we will probably call it "junk," or something equivalent in meaning to 'junk,' instead of a "poor knife," if it has no blade; though we might still call it a "handle-less knife" and not "junk"if it has a blade but no handle.

For the possession of a blade is regarded as — and is — causally more necessary for cutting, for an object's serving as a knife, than the possession of a handle.)

(2) If, as it stands, it is capable of having the associated use U — or is regarded as capable of having use U —

but in a lower degree than the

[357]

person calling it a "poor X" expects a thing called "X" to have, it being again assumed that the prevalent causal conditions are standard conditions and that "X" is intact qua "X." Taking knives again as our illustration, a knife may be said to be a poor knife if its blade, say, is rusty or is somewhat blunt, if it has a very short handle, or a very short blade, or has several or all of these defects together.

In other words, if it lacks certain characteristics which would have enabled it to have the peculiar use — cutting — associated with it qua knife, with the ease and efficiency expected of a knife; or has certain characteristics which prevent it from serving as expected.

The above analysis,
as well as further analysis,
tends to indicate
that

in such statements as
"a is a good X,"
"b is a poor X,"
where 'X' and 'Y' are common names of the kind we are dealing with in this paper,

(a)

'good'

does not refer to

any specific determinate or relatively determinate

characteristic or characteristics

common to all things ordinarily called "good X's,"

by virtue of which these things function better than is generally expected of things called "X's," or by virtue of which they are called good "X's";

(b) that,

similarly,

'poor' or 'not good' do not refer to any specific determinate or relatively determinate characteristic or characteristics present in all things called "poor Y's," which prevent them from functioning well or even functioning at all;

nor do these terms refer to any specific determinate or relatively determinate characteristic or characteristics absent from all things called "poor Y's" or "not good Y's," as a result of which they are incapable of functioning well or even functioning at all.

The determinate or relatively determinate characteristics of a knife

may vary

qualitatively and in degree,
and yet

we may still

properly call it a good knife — so long as it serves well (and/or for a relatively long time) as a knife.

Similarly a knife may be capable of cutting things it is not ordinarily expected to cut, by possessing one or more of a variety of determinate or relatively determinate characteristics of certain kinds.

The same applies to "poor knives" or "not good knives."

Of course many knives said to be good knives

may and do have

one or more determinate or relatively determinate characteristics in common, in the same way as many knives in general (say those of the same brand) whether good or not, may and do have such characteristics in common.

But our point is that

we do not restrict

the ordinary application of the term 'good' to knives which have one or more specific characteristics in common.

What determines in ordinary usage the application — and the range of application — of 'good' is something other than the possession of certain characteristics in common:

as we have been trying to show,
it is

the degree, and the nature of the conditions

under which,
a given kind of use is possessed.

On this basis, there is

no necessity,

no need

for all things called "good X's" to have one or more determinate or relatively determinate characteristics in common,

[358]

in order that they may properly be called "good X's."

Similarly, mutatis mutandis, with 'poor' or 'not good.'¹³

---

¹³ It is interesting to note that in English there is no meaning or sense of 'bad' (with the exception, perhaps, of 'bad' in 'bad food') which corresponds to the above meaning or sense of 'good.'

The antonym of 'good' in this meaning or sense is 'not good' or 'poor,' and not 'bad.'

We cannot properly say "a bad knife," "a bad game,""a bad pen," in the same way in which we speak of a "good knife,"a "good game,"a "good pen," and so on.

In other words, there does not seem to be a meaning or sense of 'bad' in English which is synonymous with 'not good' (or 'poor') in the above sense or meaning.

But there may be some languages in which we can properly say "This is a bad knife," "This is a bad game," and the like, in the sense of "This is a poor knife," "This is not a good game."

As a matter of fact, this is admissible in one Asiatic language I know, namely, in colloquial or conversational (though not in literary) Armenian.

---

One
final
word.

There is no hard and fast dividing line between cases where an object or phenomenon would be called a "poor X" and cases where it would not be called "X" at all.

Consider an object having the general structure of an armchair, but with one of its legs considerably shorter than the rest.

This object we call a poor armchair or not a good armchair (as well as, say, a "lame" armchair), rather than something else — say "junk."

The situation here is relatively straightforward, since the object in hand does serve as a separate seat for one, etc., though less satisfactorily than anything generally called a "chair" or "armchair" without qualification.

A chair is required to afford the sitter some degree of rest or comfort;
whereas this particular chair is uncomfortable.

But one can still sit upon it.

Now suppose we have a similar object, but with one of its "legs" missing.

This object cannot, as it stands, serve as a separate seat for one.

Despite that, we would still call it a chair — qualified with (say) 'broken.'

In this case, however, we would not speak of it as a "poor chair."

We call it a chair though we know that it cannot, as it stands, serve as one;

but at the same time we do not expect it to serve as a chair.

That is why we do not call it a "poor chair."

Finally, if we rip a chair apart, leaving the "seat" intact, we do not call the latter either a "poor chair," or a "broken chair," but rather, the "seat of a chair."

All this goes to show
that
the transition

from
(i) "chair" (without qualification)
to
(ii) "poor chair"
(or "not a good chair"),
and, say,
"chair with cracked seat"
to
(iii) "broken chair"
(say in the case of a chair with a leg missing),
but not "poor chair"
to
(iv) "junk,"
is not clear-cut and fixed,
but gradual and continuous and blurred.

This also goes to show that

there is no sharp demarcation line

between

"standard" conditions

for the use of a name 'X' (say 'chair')
and

"non-standard" conditions.

HAIG KHATCHADOURIAN.
American University,
Beirut.

---