American Volume

Philosophical 15, Number

Quarterly 1, January 1978

IV. TRUTH
DAVID
lived "IA7E can truly say that Sherlock Holmes * * in Baker to off liked show and that he Street, cannot that he We his mental powers. truly say in was a devoted family man, or that he worked close cooperation with the police. It would be nice if we could take such descrip? at their face value, characters tions of fictional same form to the them subject-predicate ascribing as parallel descriptions of real-life characters. Then the sentences "Holmes wears a silk top hat" and "Nixon wears a silk top hat" would both be false because the referent of the subject term?fictional
Holmes or real-life Nixon, as the case may be?

IN FICTION
LEWIS
exact about its size. Then do truly say anything we perhaps have a fictional chorus, but no fictional of this chorus and hence no number of members members? No, for we can truly say some things about the size. We are told that the sisters and
cousins, even without the aunts, number in dozens.

should not suppose that the The Meinongian in of fictional characters quantifiers descriptions range over all the things he thinks there are, both fictional and non-fictional; but he may not find it to are how the say just ranges of quantification easy
to be restricted. Consider whether we can

of lacks the property, expressed by the predicate, a silk top hat. The only difference would wearing and "Nixon" be that the subject terms "Holmes" sorts: one a have referents of radically different
fictional character, the other a real-life person of

that Holmes was more intelligent than anyone else, to before or since. It is certainly appropriate such compare him with some fictional characters, as Mycroft and Watson ; but not with others, such as Poirot or "Slipstick" Libby. It may be appro? some non-fictional priate to compare him with
characters, such as Newton and Darwin; but

truly

say

flesh and blood. that a treatment along these I dont't question be made to work. Terence lines could Meinongian to Parsons has done it.1 But it is no simple matter overcome that arise. For one thing, the difficulties is there not some perfectly good sense in which is a real-life person of flesh and like Nixon, Holmes, are stories about the exploits of blood? There from other fires and hobbits, planets, super-heroes
storms, vaporous

or others, such as Conan Doyle ably Frank Ramsey. "More than anyone intelligent else" meant like "more intelligent than something anyone else in the world of Sherlock Holmes." The inhabitants of this "world" are drawn partly from the fictional side of the Meinongian domain and from the non-fictional side, exhausting partly
neither.

not with

prob?

it would be to class the persons. But what a mistake stories with these! Unlike Clark Kent et al., Holmes is just a person?a Sherlock Holmes person of flesh and blood, a being in the very same category as
Nixon.

intelligences,

and

other

non

Finally,
about

the Meinongian
characters

must
are

tell us why
cut off,

truths

fictional

sometimes

also the problem of the chorus. We Consider can truly say that Sir Joseph Porter, K.C.B., is attended by a chorus of his sisters and his cousins and his aunts. To make this true, it seems that the
domain of fictional characters must contain not

only Sir Joseph

sisters and cousins

himself,
and

but also plenty

aunts. But how

of fictional
many?five

dozen,
the

perhaps?
numbers

No,

for we
five dozen

cannot
exactly.

truly say that

We cannot

though not always, from the consequences they lived ought to imply. We can truly say that Holmes at 221B Baker Street. I have been told2 that the only building at 22 iB Baker Street, then or now, was a bank. It does not follow, and certainly is not lived in a bank. true, that Holmes The way of the Meinongian is hard, and in this paper I shall explore a simpler alternative. Let us not take our descriptions of fictional characters at face value, but instead let us regard them as for longer sentences beginning with abbreviations
an operator "In such-and-such fiction . . .". Such a

chorus

phrase

is an intensional

operator

that may
and

be pre

1 In "A to Meinongian The Journal of Philosophy, vol. 71 (1974), pp. 561-580, Semantics," Prolegomenon 1 (1975), pp. 73-86. of Fictional Grazer Philosophische Studien, vol. gian Analysis Objects," 2 I have also been told that there has never been any building at that address. It doesn't matter which

in "A Meinon?

is correct.

37

38 fixed to a sentence to form

AMERICAN

PHILOSOPHICAL

QUARTERLY

</>

a new

sentence.

But

Holmes
and also

are not abbreviations

are not For true just instance

then the prefixed operator may be dropped by way of abbreviation, leaving us with what sounds like but differs from it in sense. the original sentence </> liked to show off, you Thus if I say that Holmes will take it that I have asserted an abbreviated "In the Sherlock of the true sentence version liked to show off." As for Holmes stories, Holmes sentence "Holmes liked to show off," the embedded taken by itself with the prefixed operator neither we may explicitly present nor tacitly understood, abandon it to the common fate of subject-predicate
sentences with denotationless subject terms: auto?

of prefixed
because these:

sentences,
is

"Holmes"

denotationless.

Holmes Holmes
resurrected.

is a fictional character. was killed off by Conan Doyle, has acquired a cultish symbolizes mankind's would
on Nixon.

but later

Holmes Holmes
for truth.

following. ceaseless

striving the

Holmes
goods

not have needed have solved

tapes to get

Holmes
sooner

could
than

the A.B.C.

murders the proper

can

matic
taste.

falsity

or lack of truth value,

Poirot.

according

to

I shall have nothing

treatment of these

to say here about

are say about Holmes Many things we might They may or may not be ambiguous. potentially
taken prefix Context, as abbreviations the Sherlock and for sentences Holmes sense "In carrying stories. will the . .". usually

sentences.

If the Meinongian

handle

content, :

common

resolve
sentences

the ambiguity

in practice.

Consider

these

lived in Baker Street. Holmes Holmes lived nearer to Paddington toWaterloo Station.
Holmes was just a person?a person

Station
of flesh

than
and

that is an them with no special dodges, over his of mine.) advantage approach of prefixing explains why truths The ambiguity are sometimes cut off characters about fictional from their seeming consequences. Suppose we have an argument (with zero or more premisses) which sense that it is impossible for is valid in the modal the premisses all to be true and the conclusion false.

blood.
Holmes really existed.

Then for many

detective

it seems clear that we obtain another valid if we prefix an operator "In the fiction argument
. ." uniformly to each premiss and to the con?

Someone
Street. London's

lived
greatest

years
in

at 22 iB Baker
used cocaine.

f.

1900

clusion of the original fiction is closed under

In/, ij?x, ...,In/,

argument. Truth implication.

*f?n

in a given

All of them are false if taken as unprefixed, simply because Holmes did not actually exist. (Or perhaps at least some of them lack truth value.) All are
true if taken as abbreviations for prefixed sentences.

But

? ifwe prefix the operator

of the original premisses

:. in/,

"In the fiction/.

and not to others,

. ."

to some

The first three would probably be taken in the latter way, hence they seem true. The rest would probably
be The taken in the former way, hence they seem false. sentence

No would
The

detective

ever solved almost

all his cases.

probably be taken as unprefixed and hence true, though it would be false if taken as prefixed.
sentence

take some but not all of the premisses as tacitly prefixed, then in general neither the original nor the prefixed conclusion "In the conclusion <f> we con? In the inference will follow. <?" fiction/, two premisses. The there were sidered earlier was at 22 Street Baker iB lived that Holmes premiss true only if taken as prefixed. The premiss that the was a bank, on only building at 221B Baker Street or if we
the other hand, was true only if taken as unpre?

Holmes

and Watson

are identical.

is sure to be taken as prefixed and hence false, but that is no refutation of systems of free logic3 which would count it as true if taken as unprefixed. to concede that some truths about (I hasten
3 For the system given in Dana instance, Scott, Schoenman (London, of the Century, ed. by Ralph "Existence 1967).

fixed ; for in the stories there was no bank there but the premisses as we rather a rooming house. Taking make them true, that in the ways naturally would : conclusion the neither follows unprefixed nothing lived in a bank nor the prefixed con that Holmes
and Description in Formal Logic" in Bertrand Russell: Philosopher

TRUTH

IN

FICTION

39 even if correct so far as it

that in the stories he lived in a bank. the unpre? both premisses as unprefixed, Taking is follows but the first premiss fixed conclusion the pre? both premisses as prefixed, false. Taking fixed conclusion follows but the second premiss is elusion r
fabe.4

rather goes.

uninformative,

Our
about such

remaining
the analysis . . .".

task is to see what may

of the operators already "In noted I have that

be said
truth

such-and

fiction

is closed under in a given fiction implication. Such closure is the earmark of an operator of rela?
tive analyzed necessity, as an a intensional restricted that may be operator over universal quantifier

A second problem arises out of an observation by Let us assume that Conan Doyle Saul Kripke.5 indeed wrote the stories as pure fiction. He just of anyone made them up. He had no knowledge nor had who did the deeds he ascribed to Holmes, he even picked up any garbled information originat? It may nevertheless be, ing in any such person. that our own world is one purely by coincidence, of the worlds where the plot of the stories is enacted.
Maybe there was a man whom Conan Doyle never

possible worlds. So we might proceed as follows: a prefixed sentence "In fiction/, is true (or, as <f>" we shall also say, <f> is true in the fiction/) is iff <f>
true at every possible world in a certain set, this set

heard of whose actual stories in every detail. "Sherlock Holmes." surely possible! Now
Holmes," used, refer as to used in the man

adventures chanced to fit the he even was named Maybe but incredible, Improbable, consider the name "Sherlock
the stories. whom Does Conan the Doyle name, never so

being As

somehow determined by the fiction/ a first approximation, we might consider of worlds is those where the the fiction exactly plot
where a course of events takes place that

heard
monymous

of? Surely
name

not!
is

It is irrelevant
used by some

that a ho
not

people,

enacted,

is true in the Sherlock the story. What matches is true at all stories would Holmes then be what of those possible worlds where there are characters who have the attributes, stand in the relations, and in the stories to do the deeds that are ascribed and the rest. (Whether these Holmes, Watson, characters would then beHolmes, Watson, and the
rest is a vexed question that we must soon con?

to refer to this man. We Doyle, including must distinguish the homonyms, between just as we would distinguish the name of London (Eng? name of London land) from the homonymous that the name It is false at our world (Ontario). Conan
"Sherlock Holmes," in the as used in the stories, refers to

someone. Yet
as used

it is true in the stories that this name,

stories, refers to someone. So we have

sider.) I think this proposal is not quite right. For one Even the thing, there is a threat of circularity.
Holmes stories, not to mention fiction written in

that is true in the stories but found something false (under our improbable supposition) at one of the worlds where the plot of the stories is enacted. it will be helpful In order to avoid this difficulty, if we do not think of a fiction in the abstract, as
a string of Rather, a fiction or sort. of that something on a is a story told by a storyteller tell his around tales occasion. He may sentences or he may type a manuscript and send

in the form of less explicit styles, are by no means chronicles. An intelligent and in? straightforward
formed reader can indeed discover the

could write it down in the form of a fully explicit chronicle if he liked. But this extraction of plot
from text is no trivial or automatic task.

plot,

and

particular the campfire

the reader
what is true

Perhaps

accomplishes
in the

it only

by figuring
is, only by

out
excer

it to his publisher, but in either case there is an act acts of storytelling, of storytelling. Different Pierre Menard re-tells different fictions. When
Quixote, that is not the same fiction as Cer?

stories?that

Don

of the very concept of cising his tacit mastery truth in fiction that we are now investigating. If so, then an analysis that starts by making uncritical use of the concept of the plot of a fiction might be
4Thus 5

even if they are in the vantes' Don Quixote?not same language and match word for word.6 (It had copied if Menard would have been different
Cervantes' fiction from memory, however; that

I have given closely follows in Fiction" that of John Heintz, "Reference and Inference far, the account (unpublished). stated in his addenda to "Naming and Necessity" and in Semantics of Natural ed. by Gilbert Harman Briefly Language, Donald Davidson a at at held at the conference lecture given (Dordrecht, 1972) ; and discussed greater length in an unpublished of Western Ontario in 1973 and on other occasions. My views and Kripke's to some extent. He also stresses University overlap I have called the ambiguity what of prefixing as engaged and regards the storyteller in pretence. The conclusions he draws from the present observation, differ greatly from mine. however, 6 "Pierre Menard, Author of the Quixote" in Ficciones New York, Jorge Luis Borges, translation, 1944; English (Buenos Aires, 1962).

40

AMERICAN

PHILOSOPHICAL

QUARTERLY

would

not have been what I call an act of story? telling at all.) One act of storytelling might, how? ever, be the telling of two different fictions : one a told to the children and the harmless fantasy
censors, the other a subversive allegory simul?

life Holmes might have had his real-life Watson who told true stories about the adventures he had witnessed. But even if his memoirs matched Conan fiction word for word Doyle's they would not be
the same stories, any more than Cervantes' Don

taneously told to the cognoscenti. is pretence. The storyteller purports Storytelling to be telling the truth about matters whereof he has to be talking about He purports knowledge. characters who are known to him, and whom he refers to, typically, of their ordinary by means proper names. But if his story is fiction, he is not his pretence has really doing these things. Usually
not the slightest tendency to deceive anyone, nor

So our world Quixote is the same story as Menard's. still not be one where the Holmes would stories? the same Holmes stories that Conan Doyle told as fiction?were told as known fact.) On the other hand, any world where the story is told as known fact rather than fiction must be among the worlds where the plot of the story is enacted. Else its enactment could be neither known nor truly told
of.

has he the slightest intent to deceive. Nevertheless he plays a false part, goes through a form of telling known fact when he is not doing so. This is most apparent when the fiction is told in the first person. to be a doctor named Conan Doyle pretended truthful memoirs of Watson, engaged in publishing events he himself had witnessed. But the case of not narrative is essentially different. third-person The author purports to be telling the truth about
matters he has somehow come to know about,

I rely on a notion of trans-world identity for of word-for-word stories; this is partly a matter match and partly a matter of trans-world identity a counterpart for acts of (or perhaps relation) storytelling. Here at our world we have a fiction/, told in an act a of storytelling; at some other world we have an act a' of telling the truth about known matters of fact; the stories told in a and a' match
word for word, and the words have the same mean?

though how he has found out about them is left unsaid. That is why there is a pragmatic paradox
akin that tale." to ends contradiction ". . . and so in a none third-person were left narrative to tell the

that mean that the other world is one ing. Does told as known fact rather than fiction? where/is as the case of Menard Not necessarily, shows. It is
also required that a and a' be the same act of

I suggest, are The worlds we should consider, the worlds where the fiction is told, but as known fact rather than fiction. The act of storytelling occurs, just as it does here at our world; but there it iswhat here it falsely purports to be : truth-telling
about matters whereof the teller has

Our own world is really a fiction

act of storytelling

cannot be such a world ; for if it that we are dealing with, then the
at our world was not what it

knowledge.7

telling (or at least counterparts). How bad is this ? like to know more about the Surely you would criteria of trans-world identity (or the counter? part relation) for acts of storytelling, and so indeed I. But I think we have enough of a would grip to on. no see make it worthwhile I threat of going here, since I see no way of using the circularity concept of truth in fiction to help with the analysis of trans-world identity of acts of storytelling. Suppose a fiction employs such names as "Sher?
lock Holmes." At those worlds where the same

story?

to be. It does not matter if, unbeknownst purported to the author, our world is one where his plot is enacted. The real-life Sherlock Holmes would not have made Conan Doyle any less of a pretender, if Conan Doyle had never heard of him. (This real

story is told as known fact rather than fiction, those names really are what to be: they here purport
ordinary to the proper storyteller. names Here of existing at our characters world, the known story?

teller only

pretends

that "Sherlock

Holmes"

has

7There are exceptions. a mixture to be uttering Sometimes the storyteller of truth and lies about matters whereo purports or ravings giving a distorted he has knowledge, reflection of the events, or the like. Tolkien to be the trans? purports explicitly an ancient come into his possession lator and editor of the Red Book of Westmarch, book that has somehow and that he some? to be a reliable how knows to be its author, record of the events. He does not purport not write else he would in English. of the Red Book by several hobbits the composition is recorded in the Red Book itself.) I should say the same about a (Indeed, historical novel written in English in which the narrator is an ancient Greek. The author does not pretend to be the first-person truthful narrator to be someone of our time who somehow but rather pretends has obtained the Greek narrator's himself, story, cases also, the thing to do is to consider knows it to be true, and passes it on to us in translation. In these exceptional those worlds where the act of storytelling to be?ravings, is whatever it purports reliable translation of a reliable source, or really at our world. cases in the remainder I shall omit mention whatever?here of these exceptional of this paper.

TRUTH

IN

FICTION

41 so not comes out true in the

the name. name, an

semantic We as have used

character no here reason at our

of at

an all

ordinary to suppose really it may

proper that the does have

consideration, fiction.

enough

world,

that character.
ordinary

As we use it, itmay

name. Indeed,

be very unlike
have a

proper

sense, governed largely by the highly non-rigid are found descriptions of Holmes and his deeds that in the stories. That is what I suggest: the sense of as we use it is such that, for "Sherlock Holmes" w stories are told as Holmes the where world any known fact rather than fiction, the name denotes at inhabitant of w it is who there plays w whichever Part ofthat role, of course, is to Holmes. of role the bear the ordinary proper name "Sherlock Holmes". But that only goes to show that "Sherlock Holmes"
is used at w as an it is so used here.8, ordinary 9 proper name, not that

For example, I claim that in the Holmes stories, Station than to lives nearer to Paddington Holmes Station. A glance at the map will show Waterloo you that his address in Baker Street ismuch nearer to Paddington. is not part of the Yet the map so as I it and far know is never stated or stories; in themselves the stories that Holmes lives implied nearer to Paddington. There are possible worlds the Holmes stories are told as known fact where rather than fiction which differ in all sorts of ways
from ours.

lives in a London
the London

Among

these

are

worlds

where

Holmes

arranged

very differently
where

from

of our world,

a London

Holmes's

a that whenever I also suggest, less confidently, world w is not one of the worlds just considered, as we use it is such the sense of "Sherlock Holmes"
as to assign it no denotation at w. That is so even

address in Baker Street ismuch closer toWaterloo Station than to Paddington. that such a distortion of (I do not suppose need the prevent otherworldly geography places
there from called being the "London," same Station,". "Paddington as, or counterparts of, that there are worlds . . their

if the plot of the fiction is enacted by inhabitants told the of w. If we are right that Conan Doyle that then it follows stories as fiction, Holmes here at our is denotationless "Sherlock Holmes" the real-life Sherlock It does not denote world. never heard of, if Holmes whom Conan Doyle such there be. We have reached a proposal I shall call Analy? sis 0: A sentence of the form "In fiction f, </>"is true iff at world true where f is told as known fact is every <f> rather than fiction. Is that right? There are some who never tire of telling us not to read anything into a fiction that is and Analysis 0 will serve to not there explicitly, of who hold this view in the those usage capture
its most that such extreme a usage form. is at I do all believe, common. Most not however, of us are

actual namesakes.
not challenge my

But if I am wrong,
claim

that still does

where

it is the stories are told as known fact but where true that Holmes lives closer toWaterloo than to For it is open to us to regard the Paddington.
place-names, as used in the stories, as fictional

names with non-rigid senses like the non-rigid sense I have already ascribed to "Sherlock Holmes." That would mean, that "Paddington incidentally,
Station," actual as used station of in that the stories, does not denote the name.)

I claim that it is true, though not Similarly, explicit, in the stories that Holmes does not have a third nostril; that he never had a case in which
the murderer turned out to be a

that he solved his cases without

revelation; Saturn; bizarre and worlds that that where he he never wears visited

purple the

gnome; moons There are told of are as

the aid of divine

content to read a fiction against a background of content well-known into" the fiction fact, "reading that is not there explicitly but that comes jointly from the explicit content and the factual back? 0 disregards the background. ground. Analysis too it into many possible worlds Thereby brings

underpants. stories the Holmes

known
false.

fact

but

where

all

of

these

things

are

Strictly speaking, it is fallacious to reason from a mixture of truth in fact and truth in fiction to con

8A rather similar treatment of fictional different from mine in that it allows the actual and purported of names, meanings to be the same, is given "Sherlock Holmes'' in Robert "Assertion" Stalnaker, (unpublished). 9 of us have never read the stories, could not produce sense of "Sher? the descriptions that largely govern the non-rigid Many sense as the most lock Holmes," in just the same is no problem There here. yet use this name expert Baker Street Irregular. causal picture of the contagion of meaning, in "Naming and Necessity" senses, Kripke's {op. cit.), will do as well for non-rigid as for rigid ones. The uses "Sherlock Holmes" sense if he has picked in its standard non-rigid it up (in the right way) ignoramus from someone who knew the governing or who picked it up from someone else who knew them, or . . .Kripke's descriptions, of rigidity could not be defended doctrines without the aid of his doctrine of contagion of meaning; without rigidity, contagion on the other hand, seems unproblematic.

42 elusions about truth in fiction.

AMERICAN

PHILOSOPHICAL

QUARTERLY

From

a mixture

of

prefixed and unprefixed premisses, nothing follows. But in practice the fallacy is often not so bad. The factual premisses inmixed reasoning may be part of the background against which we read the fiction. They may carry over into the fiction, not because in the fiction to make there is anything explicit to them true, but rather because there is nothing in the Holmes them false. There is nothing make
stories, for instance, that gives us any reason to

of the broad bracket our background knowledge outlines of London geography. Only a few details to do details having need changing?principally the stations with 221B Baker Street. To move
around, or even to regard their locations as an open

true but \fj is not true. It is vacuously true iff <f> is true at no possible worlds. omit (I accessibility restrictions for simplicity.) back to truth in fiction, recall that the Getting trouble with Analysis 0 was that it ignored back? and into consideration ground, thereby brought bizarre worlds that differed gratuitously from our actual world. A fiction will in general require some so if it is a from actuality, the more departures fantastic fiction. But we need to keep the departures
from actuality under control. It is wrong, or at

least eccentric, to read the Holmes stories as if they for all we know be taking place at a world might
where three-nostrilled detectives pursue fiction as purple

gnomes.
analyze

The
statements

true in for. What's question, would be uncalled fact about their locations is true also in the stories.
Then it is no about error what to reason else from in such the facts stories. to conclusions is true

remedy
of

is,
truth

roughly
in

speaking,
counter

to

it all before. Reasoning You've heard truth in fiction is very like counterfactual
ing. We make a supposition contrary to

about reason?
fact?

is true in the Sherlock Holmes stories factuals. What is what would be true if those stories were told as known fact rather than fiction. to my treatment of this out according Spelling
counterfactuals, we have Analysis i :A sentence of

what if this match about what would

factual match forth. situation, was dry, But we do

had been struck ? In reasoning have happened in that counter

we there use was factual factual oxygen premisses. about, and altogether The so

the form "In the fiction f, <?"is non-vacuously true iff some world where f t? told as known fact and <f> is true differs
less from our actual world, on balance, than does any

not

use

premisses

freely, since some of them would fall victim to the change that takes us from actuality to the envisaged situation. We do not use the factual counterfactual
premiss at the that time the match in a as far the question, second as we was or later. must inside that We to the matchbox it was depart a reach at room from possible comes

world where f is told as knownfact and </> is not true. It is no are true there worlds where f is iff vacuously possible told as known fact. consideration of the (I postpone
vacuous case.)

We What
as we

sometimes speak of the world of a fiction. is true in the Holmes stories is what is true,
say, "in the world of Sherlock Holmes."

That

temperature actuality where world

we speak this way should suggest that right to consider less than all the worlds where
the is enacted, stories and are less told even as known than

it is the
all fact.

counterfactual

supposition do not make

of the stories plot where the worlds

true (and that might

is a fantastic one). But

be quite
we

far if the supposition

gratuitous

as in actuality, "In the world of Sherlock Holmes," Baker Street is closer to Paddington than Station
to Waterloo Station and there are no purple

changes. We hold fixed the features of actuality that do not have to be changed as part of the least the supposition true. We disruptive way of making
can safely reason from the part have of our factual back?

gnomes.
guage to

But
the

it will
extent

not do to follow
of supposing that

ordinary
we can

lan?
some?

ground
By

that is thus held fixed.

now, several authors treated counter

factual conditionals Differences of detail

unimportant for our

along the lines just sketched. these treatments are between

present purposes. My own

of the version10 runs as follows. A counterfactual form "If it were that </>,then it would be that ^r"
is non-vacuously both (f>and world, on true iff are true than iff some differs does world possible our less from any world where where actual <f>is

how single out a single one of the worlds where the of stories are told as known fact. Is the world a world where Holmes has an Sherlock Holmes even or an odd number of hairs on his head at the is moment What when he first meets Watson? to It is absurd blood Lestrade's type? Inspector about the world of suppose that these questions
Sherlock Holmes have answers. The best explana?

tion of that is that the world*

are and plural, swers at different the questions ones. If we

of Sherlock
have may

Holmes
an? that

different assume

balance,

10Given

in Counter/actuals

(Oxford,

1973).

TRUTH

IN

FICTION

43 of concertina movement and could

some

of

the

worlds

where

the

stories

are

told

as

known fact differ are the worlds of throughout them them throughout
true at some and

least from our world, then these is true Sherlock Holmes. What is true in the stories ;what is false is false in the stories; what is
false at others is neither true nor

incapable

not

have

climbed
its victim open.12

the rope.
some other

Either
way or

the
the

snake
case

reached remains

We may well
Analysis i

look askance
then

at this reasoning.
so is Gans's

But

if

false in the stories. Any answer to the silly questions just asked would doubtless fall in the last category. It is for the same reason that the chorus of Sir Joseph Porter's sisters and cousins and aunts has no determinate size: it has different sizes at ones of the worlds of H.M.S. Pina? different
fore.11

is correct

The story never quite says that Holmes was right that the snake climbed the rope. Hence there are worlds where the Holmes stories are told as known
fact, where the snake reached the victim some other

argument.

Under
pends on

Analysis
matters

i, truth
of

in a given
fact.

fiction
I am

de?
not

therefore bungled. Pre? way, and where Holmes sumably some of these worlds differ less from ours than their rivals where Holmes was right and where
Russell's viper

contingent

that accidental thinking of the remote possibility somehow enter into properties of the fiction might are the worlds where which that determining it is a contin? fiction is told as known fact. Rather, gent matter which of those worlds differ more from ours and which less, and which (if any) differ least. That is because it is a contingent it is fact?indeed fact on which all others depend? the contingent which possible world is our actual world. To the
extent that the character of our world carries over

up a rope. Holmes's
countervailing

is capable

of

concertina

movement

infallibility,
to

of course,
our

is not a
world

resemblance

no infallible Holmes. of fictional characters provides a Psychoanalysis more important example. The critic uses (what he believes to be) little-known facts of human psychol? contains
ogy as premisses, and reasons to conclusions that

actuality;

into the worlds of Sherlock Holmes, what is true in the stories depends on what our world is like. If the stations of London had been differently located, it might have been true in the stories (and not because the stories would then have been different) that Holmes lived nearer toWaterloo Station than to Paddington Station. This contingency is all very well when truth in on well-known fiction depends facts contingent
about so far our world, as it does in the given to motivate Analysis are not well examples I. It I have is more

are far from obvious about the childhood or the state of the fictional chacter. Under adult mental i his procedure is justified. Unless counter? Analysis can be found, to consider vailing considerations worlds where the little-known fact of psychology does not hold would be to depart gratuitously from
actuality.

The
aroused

psychoanalysis
vigorous

of fictional
objections. So

characters
would

has

Gans's

if anyone cared. I shall keep neutral in argument, these quarrels, and try to provide for the needs of both sides. Analysis close to it, i, or something should capture the usage of Gans and the literary
psychoanalysts. to capture Let the us find an usage alternative of their analysis opponents. conflicting

disturbing
contingent

if truth in fiction
facts that

turns out to depend

known. In

on
an

article
movement

setting
of

forth
snakes,

little-known
Carl Gans

facts about
has argued

the
as

I shall not try to say which usage ismore conducive to appreciation of fiction and critical insight.
Suppose we psychoanalysts, our world about decide, that are contra Gans and or to the little-known irrelevant unknown truth literary facts

follows

in fiction.

In "The Adventure of the Speckled Band" Sherlock Holmes solves a murder mystery by showing that the victim has been killed by a Russell's viper that has climbed up a bell rope. What Holmes did not realize was that Russell's viper is not a constrictor. The snake is therefore
11 Heintz

But

let us not fall back to Analysis o; it is not our Let us still recognize that it is only alternative. reason to to truth in fiction from perfectly legitimate a background of well-known facts. Must they really be facts ? It seems that if little
known or unknown facts are irrelevant, then so are

he supposes that for each fiction there is a single world to be considered, but a world that is in {op. cit.) disagrees; I do not know what indeterminate. to make of an indeterminate respects I regard unless it as a superposition of world, all possible ways of the indeterminacy?or, in plainer as a set of determinate worlds that differ in the respects resolving language, in question. 12Carl "How Snakes Move," Gans, Scientific American, vol. 222 (1970), p. 93. some

44 little-known or unknown

AMERICAN

PHILOSOPHICAL

QUARTERLY

errors

in

the

body

of

shared opinion that is generally taken for fact. We think we all know that there are no purple gnomes,
but what if there really are a few, unknown to any?

themselves, living in a secluded cabin Ness ? Once we set aside the usage i, it seems clear that whatever given by Analysis purple gnomes may be hidden in odd corners of our actual world, there are still none of them in the worlds of Sherlock Holmes. We have shifted to truth in fiction as the joint product of viewing a content and of explicit generally background prevalent beliefs. Our own beliefs ? I think not. That would mean that what is true in a fiction is constantly changing. Gans might not be right yet, but he would even? error if enough tually become right about Holmes's read his that Russell's article and learned people the map of viper could not climb a rope. When Victorian London was finally forgotten, it would cease to be true that Holmes to lived nearer to to than Waterloo. the say, Paddington Strange historical scholar would be in no better position to know what was true in the fictions of his period one except near Loch
than the

is clear about such a strange situation) anything that the belief in purple gnomes does not "gener? in quite the right way, and there are ally prevail" still no purple gnomes in the worlds of Sherlock Holmes. Call a belief overt in a community at a
time less shares may iff more everyone it, and conclude, or thinks so on.13 comprises less everyone that more The shares or less it, more everyone or else we overt

proper background, are the beliefs that

in the community of origin of the fiction. that the beliefs Assume, by way of idealization, overt in the community are each possible and Then we can assign to the jointly compossible.
community a set of possible worlds, called the col?

lective belief worlds of the community, comprising those worlds where the overt beliefs all exactly come true. Only if the community is uncommonly will the actual to this set. world lucky belong the collective Indeed, the actual world determines belief worlds of the community of origin of the fiction and then drops out of the analysis. (It is of
course a contingent matter what that

What was true in a fiction when it was first told is true in it forevermore. It is our knowledge of what
is true in the fiction that may wax or wane.

ignorant

layman.

That

cannot

be

right.

The beliefs where author

factual

proper background, then, consists of the that generally prevailed in the community the fiction originated: the beliefs of the and his intended audience. And indeed the
premisses that seemed to us acceptable in

is overtly believed and what there.) We are left two sets of worlds: with the worlds where the fiction is told as known fact, and the collective belief worlds of the community of origin. The first set gives the content of the fiction ; the second gives of prevalent beliefs. the background It would be a mistake the simply to consider to both sets. Fictions usually worlds that belong
contravene at least some of the

community

is

beliefs.
are purple

I can certainly
gnomes,

tell a story in which

there are none at

community's

overt

there
our

though

reasoning about Sherlock Holmes were generally in the community believed of origin of the stories.
Everyone tions of purple knew London roughly and where the were, everyone so forth. principal disbelieved sta? in

collective overtly fiction

storytellers

belief worlds. Further, it will usually be in the community believed of origin of a that the story is not told as known fact?
seldom deceive?so none of the worlds

gnomes,

One last complication. Suppose Conan Doyle was a secret believer in purple gnomes; thinking that his belief in them was not shared by anyone else he kept it carefully to himself for fear of ridicule. In particular, he left no trace of this belief in his stories. Suppose also that some of his original secret believers readers likewise were in purple
gnomes. the time Suppose, a was in secret fact, believer that everyone in purple alive gnomes, at

each thinking that his own belief was not shared by it is clear (to the extent that anyone else. Then
13A better definition of overt

the fiction is told as known fact can be a where collective belief world of the community. Even if the two sets do overlap (the fiction is plausible and the author palms it off as fact) the worlds that belong to both sets are apt to be special in ways having to do with what is true in the fiction. nothing a of the tells in story bungled Suppose burglary recent times, and suppose it ends just as the police reach the scene. Any collective belief world of ours this story is told as known fact is a world where where covered up; the burglary was successfully for it is an overt belief among us that no such
in my Convention be found (Cambridge, or even that it will it will be knowledge, (Oxford, 1972), pp. 30-42.

Mass., 1969), pp. 52-60. That be true. See also the discussion

the name of "common may belief, under knowledge" name was unfortunate, that since there is no assurance in Stephen of "mutual Schiffer, Meaning knowledge*"

TRUTH

IN

FICTION

45 natures and in the the rest. of Macheath,

a story about I write the dragon Suppose we need is something a a like Analysis beautiful bold i, but Scrulch, princess, knight, and a not. belief It from of the the collective what is of its instance standpoint perfectly applied typical is true worlds rather than the actual world. What stylized genre, except that I never say that Scrulch stories is what would be in the Sherlock Holmes breathes fire. Does he nevertheless breathe fire in to overt of the ? the beliefs because true, according my story Perhaps so, community dragons in that sort of story do breathe fire. But the explicit content of origin, if those stories were told as known fact him breathe does not make fire. Neither rather than fiction. does 2 :A sentence background, since in actuality and according Spelling this out, we have Analysis to our beliefs there are no animals that breathe fiction f, <?" is non-vacuously true iff, form "In the of the whenever w is one of the collective belief worlds of the fire. be that (It just might analytic nothing is a dragon unless it breathes fire. But suppose I community of origin of f, then some world where f is told never called Scrulch a dragon; as knownfact and <f> is true differs less I merely endowed from theworld w, on balance, than does any world where f is told as known him with all the standard attributes dragonly is not true. It is vacuously true iff there are no If Scrulch does breathe fire except fire-breathing.) fact and <f> as told known It is is where f in worlds it is inter-fictional my story, carry-over from fact. possible by or to I true of close that offer is in what other stories. 2, it, something Analysis dragons to opponents of Gans and the literary psycho? I have spoken of Conan Doyle's Holmes stories ;
up

burglary ever hit the news. That does not make it true in the story that the burglary was covered What

treacherous Tiger Brown,

story

Polly,

analysts.

but

many

other

authors

also

have

written

Holmes

areas of I shall briefly consider two remaining sketch for and difficulty dealing with strategies them. I shall not propose improved analyses, how? I am not quite sure what ever; partly because to 2 is and make, partly because Analysis changes enough already. quite complicated I have said that truth in fiction is the joint product of two sources : the explicit content of the fiction, and a background consisting either of the facts about our world (Analysis 1) or of the beliefs overt in the community of origin (Analysis 2). Perhaps there is a third source which also contri? butes : carry-over from other truth in fiction. There
are two cases: intra-fictional and inter-fictional.

stories. These would

fictional carry-over.

have
Surely

little point without

many things are

inter
true

in these satellite stories not because of the explicit content of the satellite story itself, and not because but rather be? they are part of the background,
cause they carry over from Conan Doyle's original

if instead of asking what Holmes stories. Similarly, is true in the entire corpus of Conan Doyle's Holmes of the stories we ask what is true in "The Hound we will find doubtless Baskervilles", many things
that over are from true in that story Conan Doyle's only other by virtue Holmes of carry? stories.

In the Threepenny Opera, the principal

are a treacherous gain He crew. or They constantly for another, a streetsinger. to escape danger. shows up, sings

characters

one betray is also There the ballad of

Mack
without

the Knife,
betraying

and
anyone.

goes
Is he

about
also a

his

business him so.

even in

treacherous

fellow ?The
Real people

explicit
are not

content
so very

does not make

treacherous, and

Weimar

it was not overtly believed that Germany does not make him so they were, so background either. Yet there is a moderately good reason to say that he is treacherous: in the Threepenny Opera, that's how people are. In the worlds of the Three?
Opera, everyone put to the test proves treach?

penny

erous, the streetsinger is there along with the rest, so doubtless he too would turn out to be treach?
erous nature if we is an saw more of him. His treacherous from the intra-fictional carry-over

I turn finally to vacuous truth in impossible fictions. Let us call a fiction impossible iff there is no world where it is told as known fact rather than fiction. That might happen in either of two ways. a the plot might be impossible. First, Second, that could there be possible plot might imply nobody in a position to know or tell of the events in If a fiction is impossible in the second question. to tell it as known fact would be to know then way, its truth and tell truly something that implies that its truth could not be known; which is impossible. to all three of my analyses, anything According true in an impossible fiction. whatever is vacuously That seems entirely satisfactory if the impossibility is blatant : if we are dealing with a fantasy about the troubles of the man who squared the circle, or with the worst sort of incoherent time-travel story.
We should not expect to have a non-trivial con?

cept of truth in blatantly

haps we should expect

impossible
to have one

fiction,
only

or per?
the

under

46 to be taken too

AMERICAN

PHILOSOPHICAL

QUARTERLY unlike the will not actu?

pretence?not

are impossible possible worlds as well as the possible possible worlds. But what should we do with a fiction that is not blatantly impossible, but impossible only be? cause the author has been forgetful ? I have spoken of truth in the Sherlock Holmes stories. Strictly speaking, these (taken together) are an impossible fiction. Conan Doyle himself from contradicted
one story to another about the location of Watson's

seriously?that

there

revised

versions,

ally But
fact.

be told either as fiction or as known fact. there are worlds where they are told as fic? tion, and worlds where they are told as known

original,

Even when is not quite the original fiction impossible, there may be cases in which itwould be better to consider not truth in the original fiction but rather truth in all suitably revised versions. We
have a three-volume novel set in 1878. We learn

old war wound. Still, I do not want to say that just is true in the Holmes stories ! anything I suppose that we might proceed in two steps to is true in a venially say what impossible fiction such as the Holmes stories. First, go from the original impossible fiction to the several possible
revised versions that stay one of closest to the original.

that the hero had lunch in in the first volume on a certain day. In the third volume, it Glasgow turns out that he showed up in London that same
afternoon. In no other way does this novel purport our anal

to be a fantasy of rapid transit. The

careless. We could without vacuity

author was just

apply

Then
true,

say that what

according to

is true in the original

our analyses

iswhat
of non

is

lyses directly closest worlds

worlds sults with would

to the novel as written. Since the it is told as known fact are where
remarkable astonish means anyone?for of travel, instance, the re? our

vacuous in all of these revised truth in fiction, versions. Then nothing definite will be true in the
Holmes stories about the location of Watson's

forgetful
out a

author?who
timetable

had
of

not
the

troubled
hero's

to work

careful

movements.

Since Conan Doyle put it versions the revised different places, But at least itwill be true in the stories was wounded than in the elsewhere in various Conan Doyle put the wound wound.
never there. So no revised version

in different will differ. that Watson left big toe. places, but
put the

to apply It would be more charitable not to the original story but instead
mally revised versions that make the

the analyses to the mini?

hero's move?

will

wound
the story

in the left big toe, since that would

more than consistency demands.

change

ments feasible by in 1878. available there were ways to changes in the plot.
case perhaps truth

of travel that were the means At least, that would be best if set the times right without major There might not be, and in that
in the original version?surpris?

The revised versions, will be associated with Princeton University

like the original fiction, The acts of storytelling.

ing though
do.14

some of it may

be?is

the best we

can

Received

April

30,

igy?

14 I thank thank Kripke

who friends and colleagues the many of Learned Societies Council the American discussions. for valuable

have

given me helpful for research support.

comments Special

thanks

and I on a previous version of this paper, and Saul are due to John G. Bennett