Cognitive Linguistics Research

34

Editors
Dirk Geeraerts
René Dirven
John R. Taylor
Honorary editor
Ronald W. Langacker

2006

Mouton de Gruyter
Berlin · New York
Chapter 10
Frame semantics

Frame semantics
Charles J. Fillmore

1. Introduction

With the term ‘frame semantics’ I have in mind a research program in empirical
semantics and a descriptive framework for presenting the results of such research.
Frame semantics offers a particular way of looking at word meanings, as well as
a way of characterizing principles for creating new words and phrases, for add-
ing new meanings to words, and for assembling the meanings of elements in a
text into the total meaning of the text. By the term ‘frame’ I have in mind any
system of concepts related in such a way that to understand any one of them you
have to understand the whole structure in which it fits; when one of the things in
such a structure is introduced into a text, or into a conversation, all of the others
are automatically made available. I intend the word ‘frame’ as used here to be a
general cover term for the set of concepts variously known, in the literature on
natural language understanding, as ‘schema’, ‘script’, ‘scenario’, ‘ideational scaf-
folding’, ‘cognitive model’, or ‘folk theory’.1
Frame semantics comes out of traditions of empirical semantics rather than
formal semantics. It is most akin to ethnographic semantics, the work of the
anthropologist who moves into an alien culture and asks such questions as, ‘What
categories of experience are encoded by the members of this speech community
through the linguistic choices that they make when they talk?’ A frame semantics
outlook is not (or is not necessarily) incompatible with work and results in formal
semantics; but it differs importantly from formal semantics in emphasizing the
continuities, rather than the discontinuities, between language and experience.
The ideas I will be presenting in this paper represent not so much a genuine theory
of empirical semantics as a set of warnings about the kinds of problems such a
theory will have to deal with. If we wish, we can think of the remarks I make
as ‘pre-formal’ rather than ‘non-formalist’; I claim to be listing, and as well as I
can to be describing, phenomena which must be well understood and carefully
described before serious formal theorizing about them can become possible.
In the view I am presenting, words represent categorizations of experience, and

Originally published in 1982 in Linguistics in the Morning Calm, Linguistic Society of Korea (ed.),
111–137. Seoul: Hanshin Publishing Company. Reprinted with permission.
374 Carles. J. Fillmore

each of these categories is underlain by a motivating situation occurring against a

background of knowledge and experience. With respect to word meanings, frame
semantic research can be thought of as the effort to understand what reason a speech
community might have found for creating the category represented by the word,
and to explain the word’s meaning by presenting and clarifying that reason.
An analogy that I find helpful in distinguishing the operation and the goals
of frame semantics from those of standard views of compositional semantics is
between a grammar and a set of tools – tools like hammers and knives, but also
like clocks and shoes and pencils. To know about tools is to know what they look
like and what they are made of – the phonology and morphology, so to speak – but
it is also to know what people use them for, why people are interested in doing
the things that they use them for, and maybe even what kinds of people use them.
In this analogy, it is possible to think of a linguistic text, not as a record of ‘small
meanings’ which give the interpreter the job of assembling these into a ‘big
meaning’ (the meaning of the containing text), but rather as a record of the tools
that somebody used in carrying out a particular activity. The job of interpreting
a text, then, is analogous to the job of figuring out what activity the people had
to be engaged in who used these tools in this order.

2. A private history of the concept ‘frame’

I trace my own interest in semantic frames through my career-long interest in

lexical structure and lexical semantics. As a graduate student (at the University
of Michigan in the late fifties) I spent a lot of time exploring the co-occurrence
privileges of words, and I tried to develop distribution classes of English words
using strings of words or strings of word classes as the ‘frames’ within which I
could discover appropriate classes of mutually substitutable elements. This way of
working, standard for a long time in phonological and morphological investigations,
had been developed with particular rigor for purposes of syntactic description by
Charles Fries (Fries 1952) and played an important role in the development of
‘tagmemic formulas’ in the work of Kenneth Pike (Pike 1967), the scholars who
most directly influenced my thinking during this period. Substitutability within
the same ‘slot’ in such a ‘frame’ was subject to certain (poorly articulated) con-
ditions of meaning-preservation or structure-preservation, or sometimes merely
meaningfulness-preservation. In this conception, the ‘frame’ (with its single open
‘slot’) was considered capable of leading to the discovery of important function-
ing word classes or grammatical categories. As an example of the workings of
such a procedure, we can take the frame consisting of two complete clauses and
a gap between them, as in John is Mary’s husband – he doesn’t live with her. The
substitution in this frame of but and yet suggests that these two words have (by
 Chapter 10: Frame semantics 375

this diagnostic at least) very similar functions; insertion of moreover or however¬

suggest the existence of conjunctions functioning semantically similarly to but
and yet but requiring sentence boundaries. The conjunctions AND and OR can
meaningfully be inserted into the frame, but in each case (and in each case with
different effect) the logical or rhetorical ‘point’ of the whole utterance differs
importantly from that brought about by but or yet. In each of these cases, what
one came to know about these words was the kind of structures with which they
could occur and what function they had within those structures.
In the early sixties, together with William S.-Y. Wang and eventually D. Ter-
ence Langendoen and a number of other colleagues, I was associated with the
Project on Linguistic Analysis at the Ohio State University. My work on that
project was largely devoted to the classification of English verbs, but now not only
according to the surface-syntactic frames which were hospitable to them, but also
according to their grammatical ‘behavior’, thought of in terms of the sensitivity
of structures containing them to particular grammatical ‘transformations.’ This
project was whole-heartedly transformationalist, basing its operations at first on
the earliest work on English transformational grammar by Chomsky (1957) and
Lees (1961), and in its later stages on advances within the theory suggested by the
work of Peter Rosenbaum (Rosenbaum 1967) and the book which established the
standard working paradigm for transformationalist studies of English, Chomsky
(1965). What animated this work was the belief that discoveries in the ‘behavior’
of particular classes of words led to discoveries in the structure of the grammar
of English. This was so because it was believed that the distributional properties
of individual words discovered by this research could only be accommodated if
the grammar of the language operated under particular working principles. My
own work from this period included a small monograph on indirect object verbs
(Fillmore 1961) and a paper which pointed to the eventual recognition of the
transformational cycle as an operating principle in a formal grammar of English
(Fillmore 1963).
The project’s work on verbs was at first completely syntactic, in the sense that
what was sought was, for each verb, a full account (expressed in terms of subcat-
egorization features) of the deep structure syntactic frames which were hospitable
to it, and a full account (expressed in terms of rule features) of the various paths
or ‘transformational histories’ by which sentences containing them could be
transformed into surface sentences. The kind of work I have in mind was carried
on with much greater thoroughness by Fred Householder and his colleagues at
Indiana University (Householder et al 1964), and with extreme care and sophis-
tication by Maurice Gross and his team in Paris on the verbs and adjectives of
French (Gross 1975).
In the late sixties I began to believe that certain kinds of groupings of verbs
and classifications of clause types could be stated more meaningfully if the struc-
376 Carles. J. Fillmore

tures with which verbs were initially associated were described in terms of the
semantic roles of their associated arguments. I had become aware of certain
American and European work on dependency grammar and valence theory, and it
seemed clear to me that what was really important about a verb was its ‘semantic
valence’ (as one might call it), a description of the semantic role of its arguments.
Valence theory and dependency grammar did not assign the same classificatory
role to the ‘predicate’ (or ‘VP’) that one found in transformationalist work (see,
e.g., Tesnière 1959); the kind of semantic classifications that I needed could be
made more complete and sensible, I believed, if, instead of relying on theoreti-
cally separate kinds of distributional statements such as ‘strict subcategorization
features’ and ‘selectional features,’ one could take into account the semantic roles
of all arguments of a predication, that of the ‘subject’ being simply one of them.
Questioning, ultimately, the relevance of the assumed basic immediate-constitu-
ency cut between subject and predicate, I proposed that verbs could be seen as
basically having two kinds of features relevant to their distribution in sentences:
the first a deep-structure valence description expressed in terms of what I called
‘case frames’, the second a description in terms of rule features. What I called
‘case frames’ amounted to descriptions of predicating words that communicated
such information as the following: ‘Such-and-such a verb occurs in expressions
containing three nominals, one designating an actor who performs the act desig-
nated by the verb, one designating an object on which the actor’s act has a state-
changing influence, and one designating an object through the manipulation of
which the actor brings about the mentioned state change.’ In symbols this state-
ment could be represented as [— A P I], the letters standing for ‘Agent’, ‘Patient’
and ‘Instrument’. Actually, the kind of description I sought distinguished ‘case
frames’ as the structures in actual individual sentences in which the verbs could
appear from ‘case frame features’ as representations of the class of ‘case frames’
into which particular verbs could be inserted. In the description of ‘case frame
features’ it was possible to notice which of the ‘cases’ were obligatory, which
were optional, what selectional dependencies obtained among them, and so on
(see Fillmore 1968).
We were developing a kind of mixed syntactic-semantic valence description
of verbs, and we noticed that the separate valence patterns seemed to character-
ize semantic types of verbs, such as verbs of perception, causation, movement,
etc. Within these syntactic valence types, however, it seemed that some semantic
generalizations were lost. There seemed to be important differences between give
it to john and send it to chicago that could not be illuminated merely by showing
what syntactic rules separate give from send, just as there seemed to be semantic
commonalities between rob and steal ¬buy and sell ¬enjoy and amuse, etc., which
were lost in the syntactic class separation of these verbs.
My ultimate goal in this work in ‘case grammar’ (as the framework came
 Chapter 10: Frame semantics 377

to be called) was the development of a ‘valence dictionary’ which was to differ

importantly from the kinds of valence dictionaries appearing in Europe (e.g.,
Helbig and Schenkel 1973) by having its semantic valence taken as basic and
by having as much as possible of its syntactic valence accounted for by general
rules. (Thus, it was not thought to be necessary to explain, in individual lexical
entries, which of the arguments in a [V A P I] predication of the type described
above was to be the subject and which was to be the object, since such matters
were automatically predicted by the grammar with reference to a set of general
principles concerning the mapping from configurations of semantic cases into
configurations of grammatical relations.)
Although the concept of ‘frame’ in various fields within cognitive psychology
appears to have origins quite independent of linguistics, its use in case grammar
was continuous, in my own thinking, with the use to which I have put it in ‘frame
semantics’. In particular, I thought of each case frame as characterizing a small
abstract ‘scene’ or ‘situation’, so that to understand the semantic structure of the
verb it was necessary to understand the properties of such schematized scenes.
The scene schemata definable by the system of semantic cases (a system of
semantic role notions which I held to be maximally general and defining a mini-
mal and possibly universal repertory) was sufficient, I believed, for understanding
those aspects of the semantic structure of a verb which were linked to the verb’s
basic syntactic properties and to an understanding of the ways in which differ-
ent languages differently shaped their minimal clauses, but they were clearly
not adequate for describing with any completeness the semantic structure of the
clauses containing individual verbs.
This theory of semantic roles fell short of providing the detail needed for
semantic description; it came more and more to seem that another independent
level of role structure was needed for the semantic description of verbs in par-
ticular limited domains. One possible way of devising a fuller account of lexical
semantics is to associate some mechanism for deriving sets of truth conditions for
a clause from semantic information individually attached to given predicates; but
it seemed to me more profitable to believe that there are larger cognitive structures
capable of providing a new layer of semantic role notions in terms of which whole
domains of vocabulary could be semantically characterized.
My first attempt to describe one such cognitive structure was in a paper on
‘Verbs of judging’ (Fillmore 1971) – verbs like blame ¬accuse ¬criticize – for
which I needed to be able to imagine a kind of ‘scene schematization’ that was
essentially different from the sort associated with ‘case frames’. In devising a
framework for describing the elements in this class of verbs, I found it useful
to distinguish a person who formed or expressed some sort of judgment on the
worth or behavior of some situation or individual (and I called such a person the
Judge); a person concerning whose behavior or character it was relevant for the
378 Carles. J. Fillmore

Judge to make a judgment (I called this person the Defendant); and some situa-
tion concerning which it seemed relevant for the Judge to be making a Judgment
(and this I called simply the Situation). In terms of this framework, then, I chose
to describe accuse as a verb usable for asserting that the Judge, presupposing the
badness of the Situation, claimed that the Defendant was responsible for the Situ-
ation; I described criticize as usable for asserting that the Judge, presupposing
the Defendant’s responsibility for the Situation, presented arguments for believ-
ing that the Situation was in some way blameworthy. The details of my descrip-
tion have been ‘criticized’ (see esp. McCawley 1975), but the point remains that
we have here not just a group of individual words, but a ‘domain’ of vocabulary
whose elements somehow presuppose a schematization of human judgment and
behavior involving notions of worth, responsibility, judgment, etc., such that one
would want to say that nobody can really understand the meanings of the words
in that domain who does not understand the social institutions or the structures
of experience which they presuppose.
A second domain in which I attempted to characterize a cognitive ‘scene’ with
the same function was that of the ‘commercial event’ (see Fillmore 1977b). In
particular, I tried to show that a large and important set of English verbs could
be seen as semantically related to each other by virtue of the different ways in
which they ‘indexed’ or ‘evoked’ the same general ‘scene’. The elements of this
schematic scene included a person interested in exchanging money for goods (the
Buyer), a person interested in exchanging goods for money (the Seller), the goods
which the Buyer did or could acquire (the Goods), and the money acquired (or
sought) by the seller (the Money). Using the terms of this framework, it was then
possible to say that the verb buy focuses on the actions of the Buyer with respect
to the Goods, backgrounding the Seller and the Money; that the verb sell focuses
on the actions of the Seller with respect to the Goods, backgrounding the Buyer
and the Money; that the verb pay focuses on the actions of the Buyer with respect
to both the Money and the Seller, backgrounding the Goods, and so on, with such
verbs as spend, cost, charge, and a number of others somewhat more peripheral to
these. Again, the point of the description was to argue that nobody could be said
to know the meanings of these verbs who did not know the details of the kind of
scene which provided the background and motivation for the categories which
these words represent. Using the word ‘frame’ for the structured way in which
the scene is presented or remembered, we can say that the frame structures the
word-meanings, and that the word ‘evokes’ the frame.
The structures I have mentioned so far can be thought of as motivating the
categories speakers wish to bring into play when describing situations that
might be independent of the actual speech situation, the conversational con-
text. A second and equally important kind of framing is the framing of the
actual communication situation. When we understand a piece of language, we
 Chapter 10: Frame semantics 379

bring to the task both our ability to assign schematizations of the phases or
components of the ‘world’ that the text somehow characterizes, and our ability
to schematize the situation in which this piece of language is being produced.
We have both ‘cognitive frames’ and ‘interactional frames’, the latter having
to do with how we conceptualize what is going on between the speaker and
the hearer, or between the author and the reader. By the early seventies I had
become influenced by work on speech acts, performativity, and pragmatics in
general, and had begun contributing to this field in the form of a number of
writings on presuppositions and deixis (see, e.g., Fillmore 1975). Knowledge of
deictic categories requires an understanding of the ways in which tenses, person
marking morphemes, demonstrative categories, etc., schematize the communi-
cating situation; knowledge of illocutionary points, principles of conversational
cooperation, and routinized speech events, contribute to the full understand-
ing of most conversational exchanges. Further, knowing that a text is, say, an
obituary, a proposal of marriage, a business contract, or a folktale, provides
knowledge about how to interpret particular passages in it, how to expect the
text to develop, and how to know when it is finished. It is frequently the case
that such expectations combine with the actual material of the text to lead to the
text’s correct interpretation. And once again this is accomplished by having in
mind an abstract structure of expectations which brings with it roles, purposes,
natural or conventionalized sequences of event types, and all the rest of the
apparatus that we wish to associate with the notion of ‘frame’.
In the mid-seventies I came into contact with the work of Eleanor Rosch (Rosch
1973) and that of Brent Berlin and Paul Kay (Berlin and Kay 1969) and began
to see the importance of the notion of ‘prototype’ in understanding the nature of
human categorization. Through the work of Karl Zimmer (Zimmer 1971) and
Pamela Downing (Downing 1977) on the relevance of categorizing contexts to
principles of word-formation and, in work that reflects fruitful collaboration with
Paul Kay and George Lakoff, I began to propose descriptions of word meanings
that made use of the prototype notion. One generalization that seemed valid was
that very often the frame or background against which the meaning of a word is
defined and understood is a fairly large slice of the surrounding culture, and this
background understanding is best understood as a ‘prototype’ rather than as a
genuine body of assumptions about what the world is like. It is frequently useful,
when trying to state truth conditions for the appropriateness of predicating the
word of something, to construct a simple definition of the word, allowing the com-
plexity of fit between uses of the word and real world situations to be attributed
to the details of the prototype background frame rather than to the details of the
word’s meaning. Thus we could define an orphan as a child whose parents are no
longer living, and then understand the category as motivated against a background
of a particular kind: in this assumed background world, children depend on their
380 Carles. J. Fillmore

parents for care and guidance and parents accept the responsibility of providing
this care and guidance without question; a person without parents has a special
status, for society, only up to a particular age, because during this period a society
needs to provide some special way of providing care and instruction. The category
orphan does not have ‘built into it’ any specification of the age after which it is no
longer relevant to speak of somebody as an orphan, because that understanding
is a part of the background prototype; a boy in his twenties is generally regarded
as being able to take care of himself and to have passed the age where the main
guidance is expected to come from his family. It is that background informa-
tion which determines the fact that the word orphan would not be appropriately
used of such a boy, rather than information that is to be separately built into a
description of the word’s meaning. In the prototype situation, an orphan is seen
as somebody deserving of pity and concern; hence the point of the joke about the
young man on trial for the murder of his parents who asked the court for mercy
on the grounds that he was an orphan: the prototype scene against which society
has a reason to categorize some children as orphans does not take into account
the case in which a child orphans himself.
As a second example of a category that has to be fitted onto a background
of institutions and practices we can consider the word breakfast. To understand
this word is to understand the practice in our culture of having three meals a day,
at more or less conventionally established times of the day, and for one of these
meals to be the one which is eaten early in the day, after a period of sleep, and
for it to consist of a somewhat unique menu (the details of which can vary from
community to community). What is interesting about the word breakfast is that
each of the three conditions most typically associated with it can be independently
absent still allowing native speakers to use the word. The fact that someone can
work through the night without sleep, and then at sun-up have a meal of eggs,
toast, coffee and orange juice, and call that meal breakfast, shows clearly that the
‘post-sleep’ character of the category is not criterial; the fact that someone can
sleep through the morning, wake up at three o‘clock in the afternoon, and sit down
to a meal of eggs, toast, coffee and orange juice, and call that meal breakfast,
shows that the ‘early morning’ character of the category is also not criterial; and
lastly, the fact that a person can sleep through the night, wake up in the morning,
have cabbage soup and chocolate pie ‘for breakfast’, shows that the ‘breakfast
menu’ character of the concept is also not criterial. (This in spite of the fact that
an American restaurant that advertises its willingness to serve breakfast at any
time is referring precisely to the stereotyped breakfast ingredients.) What we
want to say, when we observe usage phenomena like that, is not that we have so
far failed to capture the true core of the word’s meaning, but rather that the word
gives us a category which can be used in many different contexts, this range of
contexts determined by the multiple aspects of its prototypic use – the use it has
 Chapter 10: Frame semantics 381

when the conditions of the background situation more or less exactly match the
defining prototype.
The descriptive framework which is in the process of evolving out of all of
the above considerations is one in which words and other linguistic forms and
categories are seen as indexing semantic or cognitive categories which are them-
selves recognized as participating in larger conceptual structures of some sort,
all of this made intelligible by knowing something about the kinds of settings or
contexts in which a community found a need to make such categories available to
its participants, the background of experiences and practices within which such
contexts could arise, the categories, the contexts, and the backgrounds themselves
all understood in terms of prototypes.

3. Further illustrations and some terminological proposals

A ‘frame’, as the notion plays a role in the description of linguistic meanings, is

a system of categories structured in accordance with some motivating context.
Some words exist in order to provide access to knowledge of such frames to the
participants in the communication process, and simultaneously serve to perform
a categorization which takes such framing for granted.
The motivating context is some body of understandings, some pattern of prac-
tices, or some history of social institutions, against which we find intelligible the
creation of a particular category in the history of the language community. The
word week-end conveys what it conveys both because of the calendric seven-day
cycle and because of a particular practice of devoting a relatively larger continu-
ous block of days within such a cycle to public work and two continuous days
to one’s private life. If we had only one ‘day of rest’ there would be no need for
the word week-end; one could simply use the name of that day. If we had three
days of work and four days of rest, then too it seems unlikely that the name for
the period devoted to one’s private life would have been given that name. (If
the work week is gradually shortened, the word week-end might stay; but it is
unlikely that the category could have developed naturally if from the start the
number of days devoted to work were shorter than the number of the remain-
ing days. An acquaintance of mine who works only on Wednesdays, pleased at
being able to enjoy ‘a long week-end’, recognizes that the word is here being
used facetiously.)
The word vegetarian means what it means, when used of people in our cul-
ture, because the category of ‘someone who eats only vegetables’ is a relevant
and interesting category only against the background of a community many or
most of whose members regularly eat meat. Notice that the word designates,
not just someone who eats plant food, but someone who eats only plant food.
382 Carles. J. Fillmore

Furthermore, it is used most appropriately for situations in which the individual

so designated avoids meat deliberately and for a purpose. The purpose might be
one of beliefs about nutrition, or it may be one of concerns for animal life; but
the word is not used (in a sentence like John is a vegetarian.) to describe people
whose diet does not include meat because they are unable to find any, or because
they cannot afford to buy it.
Occasionally one comes upon a term whose motivating context is very spe-
cific. One such is the compound flip strength, used, I am told, in the pornographic
literature business. Some publishers of pornographic novels instruct their authors
to include a certain quota of high interest words on every page, so that a potential
customer, in a bookstore, while ‘flipping’ the pages of the book, will, no matter
where he opens the book, find evidence that the book is filled with wonderful
and exciting goings-on. A book which has a high ratio of nasty words per page
has high flip strength; a book which has these words more widely distributed
has low flip strength. As I understand the word, an editor of such a publication
venture might reject a manuscript, requesting that it be returned only after its flip
strength has been raised.
With this last example, it is extremely clear that the background context is
absolutely essential to understanding the category. It is not that the conditions
for using the word cannot be stated without this background understanding (rela-
tive flip strength of novels could easily be determined by a computer), but that
the word’s meaning cannot be truly understood by someone who is unaware of
those human concerns and problems which provide the reason for the category’s
existence.
We can say that, in the process of using a language, a speaker ‘applies’ a frame
to a situation, and shows that he intends this frame to be applied by using words
recognized as grounded in such a frame. What is going on here seems to corre-
spond, within the ordinary vocabulary of a language, to lexical material in scientific
discourse that is describable as ‘theory laden’: the word phlogiston is ‘theory-
laden’; the reason it is no longer used in serious discourse is that nobody accepts
the theory within which it is a concept. That is, nobody schematizes the physical
world in a way that would give a reason to speak of part of it as phlogiston.
To illustrate the point with items from everyday language, we can consider
the words land and ground (which I have described elsewhere but cannot forego
mentioning here). The difference between these two words appears to be best
expressed by saying that land designates the dry surface of the earth as it is dis-
tinct from the sea, whereas ground designates the dry surface of the earth as it
is distinct from the air above it. The words land and ground, then, differ not so
much in what it is that they can be used to identify, but in how they situate that
thing in a larger frame. It is by our recognition of this frame contrast that we are
able to understand that a bird that ‘spends its life on the land’ is being described
 Chapter 10: Frame semantics 383

negatively as a bird that does not spend any time in water; a bird that ‘spends its
life on the ground’ is being described negatively as a bird that does not fly.
Though the details are a bit tricky, the two English words shore and coast (not
differently translatable in many languages) seem to differ from each other in that
while the shore is the boundary between land and water from the water’s point of
view, the coast is the boundary between land and water from the land’s point of
view. A trip that took four hours ‘from shore to shore’ is a trip across a body of
water; a trip that took four hours ‘from coast to coast’ is a trip across a land mass.
“We will soon reach the coast” is a natural way to say something about a journey
on land; “we will soon reach the shore” is a natural way to say something about
a sea journey. Our perception of these nuances derives from our recognition of
the different ways in which the two words schematize the world.
The Japanese adjective nurui is another example of a framing word. Although
not all Japanese-speaking informants support this judgment, enough do to make
the example worth giving. In the usage that supports my point, nurui, used to
describe the temperature of a liquid means ‘at room temperature’, but it is said
mainly of liquids that are ideally hot. Kono ocha ga nurui (this tea is lukewarm)
is an acceptable sentence in the idiolects that support my point, but kono biiru
ga nurui (this beer is lukewarm) is not. It will be noticed that the English word
lukewarm does not ‘frame’ its object in the same way. A cold liquid and a hot
liquid can both become lukewarm when left standing long enough; but only the
liquid that was supposed to be hot can be described as ‘nurui’.
A large number of framing words appear only in highly specialized contexts,
such as the term flip strength discussed earlier. The legal term decedent gives us
another example of such context specialization. According to my legal informants
(and my available law dictionaries) the word decedent is used to identify a dead
person in the context of a discussion of the inheritance of that person’s property.
(The word deceased, as in the phrase ‘the deceased’, is also limited to legal or
journalistic contexts, but it is not limited to any particular subdomain within the
law.) Another example is mufti. Mufti, in the sense it once had in the military
service, refers to ordinary clothing when worn by somebody who regularly wears
a military uniform. If we see two men wearing identical suits, we can, referring
to their clothing, say that one of them is ‘in mufti’ if that one is a military officer.
The property of being ‘in mufti’ is obviously a property that has relevance only
in the context of a military community.
Given all these examples of clear cases of terms linked to highly specific
cognitive frames, we can see that the process of understanding a text involves
retrieving or perceiving the frames evoked by the text’s lexical content and assem-
bling this kind of schematic knowledge (in some way which cannot be easily
formalized) into some sort of ‘envisionment’ of the ‘world’ of the text. If I tell
you (to be somewhat ridiculous) that the decedent while on land and in mufti
384 Carles. J. Fillmore

last weekend ate a typical breakfast and read a novel high in flip strength, you
know that I am talking about a now-dead naval officer who during the period
including last Saturday and Sunday read a pornographic novel; and you know a
few other things about the man, about how he spent his time, and about the set-
ting in which this report of his activities is given. The sentence did not give you
this information directly; you had to ‘compute’ some of it by constructing, in
your imagination, a complex context within which each of the lexically signaled
framings was motivated. We see in this way that there is a very tight connection
between lexical semantics and text semantics, or, to speak more carefully, between
lexical semantics and the process of text comprehension. The framing words in
a text reveal the multiple ways in which the speaker or author schematizes the
situation and induce the hearer to construct that envisionment of the text world
which would motivate or explain the categorization acts expressed by the lexical
choices observed in the text.
The interpreter’s envisionment of the text world assigns that world both a per-
spective and a history. A report of somebody buying something evokes the frame
of the commercial event, but sees that event, for the moment at least, from the
point of view of one of its participants. Describing somebody as being on land
locates the scene in the history of a sea voyage, by noticing that it is relevant to
describe the location in this way only if this period is seen as an interruption of
a period of sea travel. Saying that somebody is at bat locates an event as one part
of a particular baseball game. Describing coffee, in Japanese, as nurui recognizes
that it was once hot and has been allowed to ‘cool’. One knows that the coffee
is currently at room temperature, but also that it did not get that way by starting
out as iced coffee.
Sometimes the perspective which a word assigns is not a perspective on the
current scene – something that might be visible in a pictorial representation of
the scene – but is that of a much larger framework. Thus, the description of some-
one as a heretic presupposes an established religion, or a religious community
which has a well-defined notion of doctrinal correctness. In a community lacking
such beliefs or practices, the word has no purpose. Sometimes a word situates
an event in a history wider than the history of the ongoing narrative. In speaking
of locations within North America, the expressions out west and back east are
frequently used. The terms have the form they do because for a large portion of
American families the settlement history of the country traced its way from the
east coast to the west coast. European immigrants first landed on the east coast;
some of them, or some of their descendants, gradually migrated westward. The
eastern part of the country, where these immigrants or their ancestors once were,
was back east; the western part of the country, not yet reached, was out west.
The expressions are used today by people whose families did not share in this
 Chapter 10: Frame semantics 385

general westward movement themselves, but the terms recall the historical basis
of their creation.
Earlier I spoke of the notion of deep cases as offering an account of the seman-
tic aspects of single-clause predications which figured in the basic grammatical
structure of clauses. A broader view of the semantics of grammar, one which
owes a great deal to the work of Leonard Talmy (see Talmy 1980) and Ronald
Langacker (Langacker 1987), sees lexical framing providing the ‘content’ upon
which grammatical structure performs a ‘configuring’ function. Thinking in
this way, we can see that any grammatical category or pattern imposes its own
‘frame’ on the material it structures. For example, the English pluperfect can be
described as having as its role, in structuring the ‘history’ of the text world, that
of characterizing the situation at a particular time (the narrative time) as being
partly explained by the occurrence of an event or situation that occurred or existed
earlier on. The progressive aspect, in its turn, schematizes a situation as one which
is continuing or iterating across a span of time. Thus, a sentence in a narrative of
the form She had been running, a form which combines the progressive and the
pluperfect forms, can have the function of explaining why, at the narrative time
point, “she” was panting, or sweating, or tired. Thus we see that the cognitive
frames which inform and shape our understanding of language can differ greatly
in respect to their generality or specificity: a lexical verb like run can give us a
specific kind of physical activity image, while the pluperfect and the progressive
combine, each in a general and abstract way, to shape the image of running in a
way that fits the current situation and to situate the event of running both tempo-
rally and in ‘relevance’ into the ongoing history of the text world.
It is necessary to distinguish two importantly different ways in which the cog-
nitive frames we call on to help us interpret linguistic texts get introduced into
the interpretation process. On the one hand, we have cases in which the lexical
and grammatical material observable in the text ‘evokes’ the relevant frames in
the mind of the interpreter by virtue of the fact that these lexical forms or these
grammatical structures or categories exist as indices of these frames; on the other
hand, we have cases in which the interpreter assigns coherence to a text by ‘invok-
ing’ a particular interpretive frame. An extremely important difference between
frames that are evoked by material in the text and frames that are invoked by the
interpreter is that in the latter case an ‘outsider’ has no reason to suspect, beyond
a general sense of irrelevance or pointlessness in the text, that anything is miss-
ing. To repeat an example that I have used elsewhere, a Japanese personal letter in
the traditional style is supposed to begin with a comment on the current season.
Somebody who knows this tradition is able to sense the relevance of an opening
sentence in a letter which speaks of the garden floor covered with leaves. The
kind of understanding which allows such an interpretation comes from outside
of the text itself.
386 Carles. J. Fillmore

Invoked frames can come from general knowledge, knowledge that exists
independently of the text at hand, or from the ongoing text itself.

4. Frame-semantic formulations of empirical semantic observations

In this section I examine a number of observations about lexical meaning or text

interpretation which permit formulations in terms of notions from frame seman-
tics. In the following section I examine a number of traditional topics in standard
semantic theorizing and raise questions about the importance they would be given
in an account of linguistic meaning of the sort we have been exploring.

4.1 Polysemy arising from alternative framings of the same lexical item
For many instances of polysemy it is possible to say that a given lexical item prop-
erly fits either of two different cognitive frames. One possibility is that a word
has a general use in the everyday language but has been given a separate use in
technical language. For example, we might wish to say that the English word
angle is understood in connection with a perceptual frame as a figure made by
two lines joined at a point in a way suggested by a bent stick. Presented in terms
of a competing procedural frame, an angle is thought of in terms of the rotation
of a line about a point, the angle itself visually represented as the line before and
after its rotation. In the procedural frame the notion of a 180 degree angle is intel-
ligible, as is the notion of a 360 degree angle. Within the perceptual frame such
notions do not fit. (The example is from Arnheim 1969: 182f.)

4.2. Alternate framings of a single situation

From a frame semantics point of view, it is frequently possible to show that the
same ‘facts’ can be presented within different framings, framings which make
them out as different ‘facts’. Somebody who shows an unwillingness to give out
money in a particular situation might be described by one person as stingy (in
which case the behavior is contrasted with being generous), and by another as
thrifty (in which case a contrast is made with being wasteful). The speaker who
applies the stingy: generous contrast to a way of behaving assumes that it is to be
evaluated with respect to the behaver’s treatment of fellow humans; whereas the
speaker who evaluates the behavior by applying to it a thifty: wasteful contrast
assumes that what is most important is a measure of the skill or wisdom displayed
in the use of money or other resources.
 Chapter 10: Frame semantics 387

4.3. ‘Contrast within frames’ versus ‘contrast across frames’

The fact that a single situation can be ‘framed’ in contrasting ways makes pos-
sible two ways of presenting a negation or an opposition. Using the contrasts
introduced in the last paragraph, if I say of somebody, He’s not stingy – he’s
really generous, I have accepted the scale by which you choose to measure
him, and I inform you that in my opinion your application of this scale was in
error. If on the other hand I say He’s not stingy – he’s thrifty, what I am doing is
proposing that the behavior in question is not to be evaluated along the stingy:
generous dimension but along the thifty: wasteful dimension. In the first case
I have argued for a particular standard in the application of an accepted scale;
in the second case my utterance argues for the irrelevance of one scale and the
appropriateness of another.

4.4. Word sense creation by frame borrowing

When a speaker wishes to talk about something for which an appropriate cogni-
tive frame has not been established, or for which he wishes to introduce a novel
schematization, he can sometimes accomplish this by transferring the linguistic
material associated with a frame which makes the distinctions he’s interested in
onto the new situation, relying on the interpreter to see the appropriateness of the
transfer. Certain new senses of words can be best understood as having originated
in this way; we might expect that such was the case in the importation of the term
bachelor into the terminology appropriate to fur seal society, to use the example
made common in lexical semantics discussion from the reminder, in Katz and
Fodor (1963), of the use of the word bachelor to designate ‘a male fur seal without
a mate during the mating season’. Lakoff and Johnson (1980) have made us aware
of the value of metaphor in conceptualization and communication, making the
persuasive case that in a great many domains of experience metaphors provide
us with the only way of communicating about those experiences.

4.5. Reframing a lexical set

Various kinds of semantic change can be illuminated by considering the phenom-
ena in frame semantic terms. One important type of change consists in reconstitut-
ing the motivating circumstances while preserving the lexical item and its basic
fit with the associated scene. People observing certain usages of English with an
eye to feminist concerns have noticed tendencies on the part of many speakers
to have certain asymmetries in the sets of conditions for using the words in the
proportion boy: man :: girl: woman. In particular, in the usage pattern that I
have in mind, males appeared to be classified as men at an earlier age than that
388 Carles. J. Fillmore

at which females are classified as woman. A number of people, sensing that this
usage pattern revealed attitudes toward females (or a history of attitudes toward
females reflected in current conventional usage possibly in independence of the
user’s own attitudes) which ought to be corrected. A number of speakers have
succeeded in modifying their usage in a way which established the age bound-
ary between the boy to man transition at the same place as that between the girl
to woman transition. The semantic change in this case is a real one, which needs
to be explained. But it would not be satisfying to see the explanation solely in
changes of the meaning of the words girl and woman; the full explanation must
assign the change to the underlying schematization on the part of the language
user. The realities (of people of both sexes getting older) have not changed, nor
have the available choices of linguistic material; what has changed (in some
speakers) is the underlying schematization, the circumstances motivating the
category contrasts.

4.6. Relexicalizing unchanged frames

A second kind of semantic change, which oddly can be illustrated with the same
words, is one in which the links between words and their frames are changed,
but the underlying schematization remains unchanged. The effort to respond to
society’s new sensitivity to the connections between language and attitudes is
perhaps easiest to manage in the short run if it does not require something as
deeply cognitive as a reschematization of the domain. A superficial rule-of-thumb
for bringing about the appearance of a raised consciousness in the realm of lan-
guage and sexism is a mechanical principle like “Where I am inclined to say girl
I should instead say woman ”. A person who adopts this rule may find that in
most cases it performs very well; but one sometimes finds oneself trapped – as
in the experience of an acquaintance of mine – when talking about very young
females; my friend found himself, several times, using the word woman when
talking about an eight-year-old girl. The fact that this friend would never acci-
dentally use the word man when talking about an eight-year-old boy shows that
the change in question is not of the reschematization type discussed in the pre-
vious paragraph. An equally clear example of the same phenomenon (as I have
discussed elsewhere – Fillmore 1972) is in the use of the word suspect where the
speaker or writer might have been inclined to use such a word as burglar, mur-
derer, arsonist, or more generally, culprit. Conscious of the legal doctrine that a
person is to be considered innocent until proven guilty, and conscious too of the
danger of committing libel, journalists and police officers have learned to iden-
tify persons accused of crimes but not (yet) legally held to be guilty of them as
suspects. A change in usage which would clearly reflect the adoption of the legal
doctrine mentioned above about guilt and innocence as the underlying cognitive
 Chapter 10: Frame semantics 389

frame would not result in some of the frequent mistakes people make in the use
of the word suspect. The word suspect is supposed to be used of a person who
is suspected of committing the crime in question; for it to be used appropriately,
there has to be some specific person of whom it can be said that that person is
suspected by someone of committing the crime. The current journalistic use of
suspect even when nobody has been accused of the crime shows that the change
is of the superficial kind, following the application of a rule of thumb that says,
“Wherever I am inclined to say culprit (etc.), I should instead say suspect.” I
have in mind such usages as can be found in reports like “Police investigating
the murder have found no clues as to the identity of the suspect.”

4.7. Miscommunication by frame conflict

The law provides many contexts in which specific new framings need to be con-
structed for familiar words. The notion innocent mentioned above is an example.
In both everyday language and legal language there is a contradictory opposition
between innocent and guilty. In everyday language, the difference depends on
whether the individual in question did or did not commit the crime in question.
In legal language, by contrast, the difference depends on whether the individual
in question has or has not been declared guilty by the court as a result of legal
action within the criminal justice system. This disparity of schematization is
responsible for frequent misunderstandings in the use of these words. An example
of such misunderstandings (which I have discussed in Fillmore 1978) was in a
conversation between a prospective juror and lawyers in a voir dire hearing in a
municipal court in Berkeley. The attorney for the defense asked the prospective
juror “Do you accept the American legal doctrine that a man is innocent until
proven guilty?” The citizen answered that a person should be treated as inno-
cent until proven guilty, but that it would be strange to say that he was actually
innocent. The attorney asked again, saying, “I’m talking about the doctrine that a
man is innocent until proven guilty. Do you or do you not accept that doctrine?”
The citizen answered that if the man is innocent, then there is no need for a trial.
(This rude answer excused the man from jury duty.) This little bit of miscommu-
nicating could easily have been avoided. The citizen was not really being asked
whether or not he accepted a particular legal doctrine, but whether or not he was
willing to adopt for the purpose of discussion in the trial which was about to start
the framing of the words innocent and guilty provided by the criminal justice
institutions in place of the everyday use of these same words.
390 Carles. J. Fillmore

4.8. Reformulations in technical language

Legal contexts give us further ways of seeing changes between general and spe-
cial-purpose framings of words. In many cases this is because the everyday sense
of a word does not cover all cases in which it should be appropriate to use the
word. In the prototype case of events fitting the word murder, one person (A),
intending to kill a second person (B), acts in such a way as to cause that person
to die. This prototype does not cover a case in which A, intending to kill B, aims
his gun at B, and kills C (who is standing next to B) instead. Some of the prop-
erties of murder relate A and B; others relate A to C. The question somebody
needs to answer, of course, is whether, for the purposes of the law, it is proper
to say that A murdered C. The law does this, not by modifying the definition of
murder so that it will cover this ‘wrong-target’ case, but by adding to the system
of legal semantics a statutory interpretation principle called ‘Transfer of Intent’
according to which A’s intent to kill B is fictitiously transferred to C so that the
definition of murder can fully fit what A did to C. With respect to judgments
of reprehensibility and legal provisions for punishment, A’s killing of C should
be treated in the same way as A’s successful killing of B would have been. The
Transfer of Intent principle makes it possible for the non-prototypic case to fall
under the same definition.
Other such reinterpretations in the law are equally founded on intentions
associated with the prototypical case. The concept of forcible entry involves one
person gaining entry to another person’s property by overcoming the resistance
of persons trying to prevent that person’s entry. The usual definition of forcible
entry, however, includes not only the situation in which the intruder physically
overpowers the other, but also the situation in which, as it is usually put, “resis-
tance would be unavailing”. If you, being twice my size and strength, insist on
being admitted to my apartment, and I meekly let you enter (on the reasonable
grounds that if we had a fight, I would lose), then too you can be charged with
forcible entry. A third example is oral agreement. Basically an oral agreement
is a contract or agreement which two parties entered into orally, that is, without
putting the agreement in a written form and without signing our names to it.
The importance of the notion oral agreement in the law is that the conditions of
its authenticity and its bindingness distinguish it from agreements that are fully
written out and signed. The critical difference, for the given legal purposes, is the
presence or absence of the signatures of the principals. The important part of the
contrast, then, is that between being signed and not being signed. Accordingly,
provisions made in the law for oral agreements also apply to written agreements
which happen not to be signed. The prototype background in which the notion
oral agreement is motivated, is one in which agreements are either made by word
of mouth or by means of documents which are written and signed. In situations
 Chapter 10: Frame semantics 391

which depart from the prototype the law has needed to determine which aspect
of the prototype contrast is legally the most salient (the presence or absence of
the signatures supporting a written document) and let that be the criterion which
specifies the contrast.

4.9. Frames for evaluation

One important area in which semantic interpretation depends crucially on lexi-
cal framing is that of attributions of value. Evaluative adjectives can contain in
their meanings reference to the dimensions, scales, or standards according to
which something is evaluated, as with adjectives like fragrant, tasty, efficient,
intelligent, etc. In many cases, however, an adjective is abstractly evaluative (as
with the English words good and bad) and interpretations of their attributive use
depend on knowledge of the ideational frames to which they are indexed. The
fact that speakers of English are able to interpret such phrases as a good pencil,
good coffee, a good mother, a good pilot, etc., shows that they are able to call
into their consciousness for this purpose the fact that a pencil is used for writing
and can be evaluated for how easy or efficient it is to write with it, or how clearly
its traces appear on the paper, the fact that coffee is a drink and can be evalu-
ated for its taste, its contribution to the drinker’s alertness, etc., that mothers and
pilots do what they professionally and conventionally do and can be evaluated for
how easily, how effectively, and how efficiently they do it. The point was made
earlier that cognitive frames called on to assist in text interpretation may derive
from general background knowledge or may be brought into play by the textual
context. This is particularly true in the case of the interpretation of evaluative
adjectives, since some nouns have frames associated with them whose evaluative
dimensions are provided in advance, while others designate things that could be
evaluated only if the context provided some basis for the evaluation. When we
come across the phrase a good stick we expect to find in the context some expla-
nation of a situation within which one stick could function better than another (for
propping a window open, for repelling a raccoon, for skewering marshmallows,
etc.). A general concept of ‘framing’ involves contextualizing or situating events
in the broadest sense possible; within linguistic semantics proper the concern is
with patterns of framing that are already established and which are specifically
associated with given lexical items or grammatical categories.

4.10. Script evocation

I said earlier about cognitive frames that to speak of one of its elements is to speak
of the others at the same time. More carefully put, to speak of one part of a frame
is to bring to consciousness, or to raise into question, its other components. This
392 Carles. J. Fillmore

effect is particularly striking in connection with the kinds of frames known as

‘scripts’, frames whose elements are sequenced types of events. Text understanding
that makes use of scriptal knowledge (on which see Schank and Abelson 1977)
involves the activation of whole-scale scripting of events on the presentation of an
event that can be seen to part of such a script. Thus, in a textlet like “He pushed
against the door. The room was empty.” we make the two sentences cohere by
assuming that the goal somebody might have in pushing against a door is to get
that door open, and that if one succeeded in getting the door open by such an act,
one could then be in a position to notice whether the room was empty. Reading
between the lines, we expand the text to mean: “He pushed against the door. the
door opened. he looked inside. he saw that The room was empty.”

4.11. Frames for texts

Discussion of text structure on the part of Robert Longacre and others shows that
languages or cultures can differ with respect to the ways in which texts with par-
ticular communicative goals can have particular conventionalized forms. Recipes
in English make consistent use of imperatives. In Hungarian recipes, first person
plural descriptions are the norm. And Longacre has described (in conversation) a
language lacking in procedural discourse uses narrative form for such purposes.
Here it would be difficult to believe that languages differ from each other in the
presence of material usable for particular kinds of discourse, it seems rather to
be the case that traditions of language use within the culture develop in different
ways in texts with different communicative goals.

5. Frame-semantic formulations of issues in technical semantics

In this section I examine a small number of topics that one traditionally finds in
standard treatises on technical semantics: proportionality, paradigms, taxonomies,
syncategorematicity, the supposed contrast between ‘dictionary’ and ‘encyclope-
dia’, the goal of descriptive simplicity and redundancy elimination, and, lastly,
the troubled notion of ‘lexical presupposition’.

5.1. Proportionality
One of the most frequently used heuristic devices for discovering and demonstrat-
ing the existence of semantic features in the vocabulary of a language is that of
setting up a proportionality involving four words and asking for intuitive agree-
ment about the identity of pairwise differences among them. Believing that man
is to woman as boy is to girl, we set up the ratio man: woman :: boy: girl. Others
 Chapter 10: Frame semantics 393

frequently used are come: go :: bring: take, look: see :: glance: glimpse, inhale:
exhale :: sniff: snort, and man: woman :: bachelor: spinster. The approach which
sees the basic semantic relations as holding among words taken in isolation fails
to help us become aware of the possibly quite separate ways in which individual
members of these proportions are fitted onto, or frame, their reality. I have already
pointed out that in many people’s speech the differentiating criterion for boy vs.
man might be importantly different from that for girl vs. woman; bring is separate
enough in its semantics from come for it to have acquired quite separate patterns
of dialect variation; and the motivation for the categories bachelor and spinster
appear to be considerably different, in spite of one’s inclination, as a systematizer,
to put the two words together. One might wish to propose that the abstract struc-
tural patterns underlying these word groups are simple and straightforward, in
the ways suggested by the proportions, even though certain facts about the world
make the domain look less orderly. I think such a proposal is not helpful, because
it is not one which asks the analyst to look for the background and motivating
situations which separately give reasons for the existence of the individual cat-
egories, one by one.

5.2. Paradigms
A prime example of semantic structure among lexical items is the ‘paradigm’;
and the best example of a lexical-semantic paradigm is the kind of display of
livestock terms represented by Table 1.

Table 1.

cattle sheep horse swine

cow ewe mare sow
bull ram stallion boar
steer wether gelding barrow

Here the proposal that we have a closed system of terms tied together by such
features as General, Female, Male, and Neuter, cross-cut by features identifying
species (Bovine, Ovine, Equine, Porcine), seems very attractive. Unfortunately
the display disguises many facts about both these words and the domain which
they appear to cover. cattle and swine are plurals; sheep and horse are not. The
words wether and barrow are known only to specialists. In the case of cattle,
cow and bull appear to have the status of ‘basic level objects’ (in the sense of
Rosch 1973), whereas the general terms have that function in the case of sheep
394 Carles. J. Fillmore

and horse. In the case of swine, a word not in the table, namely pig, is the best
candidate for ‘basic level object’ status.
In short, the regularities apparent in the paradigm (and this set of terms –
together with terms for young, newborn, etc. – make up what is generally accepted
as the best example of a semantic paradigm) are misleading. To which we ought
to add the Neuter category of the words in the bottom row is not just a ‘neutral’
category operating in the same line of business as the categories Female and Male.
The category is differently motivated in the different species, which is another
way of saying that one has different reasons for castrating a bull and a horse, one
might do it at different (relative) ages, etc,

5.3. Taxonomies
The next most common kind of lexical semantic formal structure is the ‘semantic
taxonomy’, a semantic network founded on the relation ‘is a kind of’. Scientific
taxonomies have obvious uses in scientific discourse, and research that has led to
the uncovering of folk taxonomies has been among the most important empirical
semantic research yet done. But there are two aspects of taxonomic structures that
argue against regarding them as representing merely a formal system of relation-
ships founded on a single clear semantic relation. The first is that at different levels
in a taxonomy the community might have had different reasons for introducing
the categories; the second is that the usual tree-form display of the elements of a
taxonomy does not show how it is that particular elements in the taxonomy are
‘cognitively privileged categories’ in important ways. Both of these points can be
illustrated with a ‘path’ in a taxonomy of zoological terms in English, namely
animal
vertebrate
mammal
dog
retriever

Of this set of words, dog and animal seem to be the cognitively privileged cat-
egories, privileged in the sense that they are the words that would most ordinar-
ily be used when in everyday natural talk one is describing one’s experiences.
vertebrate and mammal are terms whose employment fits a particular kind of
interactional or contextual schema (that of scientific discourse),while retriever as
a category occurs most naturally as an answer to a question about what kind of
a dog one has. Suppose that you, hearing a splash in my back yard, were to ask
me what that noise was, and suppose the fact is that my pet retriever fell in the
family swimming pool. As a way of explaining the source of the noise, it would
be natural for me to say “An animal fell in the pool” or “A dog fell in the pool”,
 Chapter 10: Frame semantics 395

but it would be very unnatural for me to say “A vertebrate fell in the pool” or
“A mammal fell in the pool”, and unnatural in a different way for me to say “A
retriever fell in the pool”. The latter three terms seem to appear more natural in
utterances used in acts of classifying, but seem unnatural when used in acts of
referring. This functional difference is not revealed within the logic of a standard
taxonomic tree.

5.4. Syncategorematic terms

It has frequently been discussed (e.g., Austin 1964, Lecture VII) that a word like
imitation does not semantically modify a word it grammatically modifies in the
standard ‘set intersection’ way. Rather, it combines with the meaning of its partner
to form a fairly complex concept. Something correctly described as imitation cof-
fee looks like coffee and tastes like coffee, and it looks and tastes like coffee not
by accident, but because somebody manufactured it so that it would have these
properties; but, whatever it is, it is not made of coffee beans. Understanding the
category, in fact, requires understanding the role of coffee in our lives and (per-
haps) the reasons someone might have for making a coffee substitute.
By contrast a word like real appears to contribute nothing at all to the noun
to which it is attached as a modifier. To describe something as real coffee is to
do nothing more than to assert that something is coffee, against the background
of (the possibility of) somebody’s suspicion that it is imitation coffee. As with
imitation, a part of a full understanding of an expression with real is knowing the
reasons one might have for providing substitutes for the thing in question. The
notion real coffee makes sense to us because we know that in some settings cof-
fee is scarce, and we know that some people find coffee damaging to their health
or held offensive by their religion. We can understand a category like real gold
or real diamond because we can imagine a reason why somebody might choose
to produce fake gold or fake diamonds, and we can imagine why someone might
have doubts about the authenticity of particular samples. By contrast, a notion
like real pants is unintelligible, because it is impossible to imagine something
looking like pants and functioning like pants which do not, by virtue of those
properties alone, count as being genuine pants.

5.5. Redundancy elimination

A common goal in structural semantics is the elimination or minimization of
redundant information in the semantic description of lexical items. Frequently a
semantic theorist will declare that the goal of a ‘semantic dictionary’ is that of
saying just enough about each word in the language to guarantee that it is semanti-
cally in contrast with each other word in the language (Bendix 1966). It is a goal
396 Carles. J. Fillmore

which presupposes the analyst’s ability to have an overview of the entire lexical
repertory of the language. Such a goal is completely antithetical to the goals of
frame semantics, since frame semantics aims at discovering what categorizing
functions the word serves in the contexts in which its use is motivated. This kind
of knowledge is in principle attainable independently of knowledge about other
words in the language, except for those relatively few cases in which the ‘mosaic’
image is appropriate, the image by which the meaning given to any one word is
dependent on the meanings of its neighboring words (as in Trier 1931).

5.6. Dictionary vs. encyclopedia

The various structuralist approaches that find a goal of redundancy elimination
relevant, also find it intelligible to draw a clear distinction between ‘dictionaries’
and ‘encyclopedias’. In particular, certain scholars insist on a distinction between
purely semantic information about words and encyclopedic information about the
designata of words. Somebody holding this view might expect to be able to jus-
tify certain characteristics of carpenters (or the concept carpenter) as belonging
to the semantic category of the noun, other distinct characteristics of carpenters
as simply being true of the individuals who satisfy the criteria associated with
the category. A frame-semantic approach would rather say that communities of
men contain individuals who by trade make things out of wood, using particular
kinds of tools, etc., etc., and would note that these people are called carpenters.
The possibility of separating some features of a full description of what carpen-
ters do as related to the concept and others as related to the people does not seem
important. There is a distinction to be made between knowledge about words and
knowledge about things, but it is not to be made in a way that serves the interests
of the semanticists I have just been describing. True ‘encyclopedic’ information
about carpenters as people might say something about wages, union affiliations,
job related diseases, etc.; such information is not a matter of dispute.

5.7. Simplicity of description

While in respect to redundancy elimination it has appeared that standard approaches
value simplicity and frame-semantic approaches do not, there is another sense in
which simplicity of description is enhanced by the frame semantics approach. A
recent lively discussion between Paul Kay and Linda Coleman on the one hand
(Coleman and Kay 1981) and Eve Sweetser on the other hand (Sweetser 1981)
concerns the possibility of a prototype background of assumptions (or, as Sweetser
calls it, a ‘folk theory’) as providing the grounding for a simplified definition of
the noun lie. On the Kay/ Coleman account, a lie is something which is (1) false
in fact, (2) believed by the speaker to be false, and (3) said in order to deceive.
 Chapter 10: Frame semantics 397

Sweetser’s suggestion is that if we can characterize a folk theory of human com-

munication involving cooperation, expressing what one believes, etc., then it is
possible to describe a lie as simply a ‘false statement’, those other understandings
we have about the concept falling out through an understanding of why one would
bother to produce a false statement.

5.8. Presupposition
Claims about ‘presuppositional’ information being associated with individual
lexical items have not received a good press. I find that within frame semantics,
the concept of lexical presupposition does not seem unjustified. Consider the case
of a verb like English chase, a verb for which a lexical presuppositionist might be
inclined to say that when it is used of two beings moving in the same course, the
movement of the one in front is presupposed, independently of whether the move-
ment of the individual designated by the subject of the verb is asserted, denied,
questioned, or supposed. In a setting in which one person is running, especially
where it is understood that that person is fleeing, it is relevant to consider whether
some other person is or is not going to try to prevent that first person from getting
away. (My illustration is with people, but that’s not an important condition.) The
verb chase exists as a category by recognition of such relevance. If I ask, “Did
anybody chase him?”, or if I say “We didn’t chase him”, our reason for understand-
ing that ‘he’ was running (fleeing) is that we know the kind of situation against
which the category chase has a reason for being. It is in that sense, it seems to
me, that one can talk about lexical presuppostions.

6. Concluding remarks

In this paper I have argued for a view of the description of meaning-bearing ele-
ments in a language according to which words (etc.) come into being only for a
reason, that reason being anchored in human experiences and human institutions.
In this view, the only way in which people can truly be said to understand the use
to which these meaning-bearing elements are being put in actual utterances is to
understand those experiences and institutions and to know why such experiences
and institutions gave people reasons to create the categories expressed by the
words. The semanticist’s job is to tease out the precise nature of the relationship
between the word and the category, and the precise nature of the relationships
between the category and the background. I believe that some of the examples I
have offered have shown the advantages of looking at language in this way.
398 Carles. J. Fillmore

Note
1. For a recent attempt to differentiate these terms, see Beaugrande 1981: 303.

References
Arnheim, Rudolf
1969 Visual Thinking. Berkeley: University of California Press.
Austin, John L.
1964 Sense and Sensibilia. (Reconstructed from manuscript notes by Geoffrey J.
Warnock.) Oxford University Press.
Beaugrande, Robert de
1981 Design criteria for process models of reading. Reading Research Quarterly
16(2): 261–315.
Bendix, Edward H.
1966 Componential Analysis of General Vocabulary: The Semantic Structure of a
Set of Verbs in English, Hindi and Japanese. Bloomington: Indiana University
Press.
Berlin, Brent and Paul Kay
1969 Basic Color Terms. Berkeley: University of California Press.
Chomsky, Noam A.
1957 Syntactic Structures. The Hague: Mouton
1965 Aspects of the Theory of Syntax. Cambridge, MA: M.I.T. Press
Coleman, Linda and Paul Kay
1981 Prototype semantics. Language 57: 26–44.
Downing, Pamela
1977 On the creation and use of English compound nouns. Language 53:
810–842.
Fillmore, Charles J.
1961 Indirect Object Constructions in English and the Ordering of Transforma-
tions. The Hague: Mouton.
1963 The position of embedding transformations in a grammar. Word 19:
208–301.
1968 The case for case. In Universals in Linguistic Theory, Emmon Bach and Rich-
ard Harms (eds.), 1–90. New York: Holt, Rinehart & Winston.
1971 Verbs of judging: An exercise in semantic description. In Studies in Linguistic
Semantics, Charles J. Fillmore and D. Terrence Langendoen (eds.), 272–289.
New York: Holt, Rinehart and Winston.
1972 On generativity. In The Goals of Linguistic Theory, Stanley Peters (ed.), 1–19.
Englewood Cliffs: Prentice Hall.
1975 Santa Cruz Lectures on Deixis. Bloomington: Indiana University Linguistics
Club.
1977a The case for case reopened. In Syntax and Semantics 8: Grammatical Relations,
Peter Cole and Jerry Sadock, (eds.), 59–82. New York: Academic Press.
 Chapter 10: Frame semantics 399

1977b Topics in lexical semantics. In Current Issues in Linguistic Theory, Roger W.

Cole (ed.), 76–138. Bloomington: Indiana University Press.
1978 On the organization of semantic information in the lexicon. In Papers from
the Parasession on the Lexicon, 1–11. Chicago: The Chicago Linguistic
Society.
Fries, Charles C.
1952 The Structure of English. New York: Harcourt, Brace & World.
Gross, Maurice
1975 Méthodes en syntaxe. Paris: Hermann.
Helbig, Gerhard and Wolfgang Schenkel
1973 Wörterbuch zur Valenz und Distribution deutscher Verben. Leipzig: VEB
Verlag Enzyklopädie.
Householder, Fred W., et al.
1964 Linguistic Analysis of English, Final Report on NSF Grant No. GS-108.
Katz, Jerrold J. and Jerry A. Fodor
1963 The structure of a semantic theory. Language 39: 170–210.
Lakoff, George and Mark Johnson
1980 Metaphors We Live By. Chicago: University of Chicago Press.
Langacker, Ronald W.
1987 Foundations of Cognitive Grammar. vol. 1: Theoretical Prerequisites. Stan-
ford: Stanford University Press.
Lees, Robert B.
1960 The Grammar of English Nominalizations. The Hague: Mouton
McCawley, James D.
1975 Verbs of bitching. In Contemporary Research in Philosophical Logic and Lin-
guistic Semantics, David Hockney et al. (eds.), 313–332. Dordrecht: Reidel.
Pike, Kenneth L.
1967 Language in Relation to a Unified Theory of the Structure of Human Behavior.
The Hague: Mouton.
Rosch, Eleanor H.
1973 On the internal structure of perceptual and semantic categories. In Cogni-
tive Development and the Acquisition of Language, Timothy E. Moore (ed.),
111–144. New York: Academic Press.
Rosenbaum, Peter S.
1967 The Grammar of English Predicate Complement Constructions. Cambridge,
MA: MIT Press
Schank, Roger C. and Robert P. Abelson
1977 Scripts, Plans, Goals and Understanding: An Inquiry into Human Knowledge
Structures. Hillsdale, N.J.: Lawrence Erlbaum.
Sweetser, Eve E.
1981 The definition of lie: An examination of the folk theories underlying a semantic
prototype. Unpublished ms.
Talmy, Leonard
1980 Grammar and cognition. Unpublished ms. University of California at San
Diego, Cognitive Science Program.
400 Carles. J. Fillmore

Tesnière, Lucien
1959 Elements de syntaxe structurale. Paris: Klingksieck.
Trier, Jost
1931 Der deutsche Wortschatz im Sinnbezirk des Verstandes. Heidelberg.
Wilson, Deirdre
1975 Presuppositions and Non-Truth-Conditional Semantics. London: Academic
Press.
Zimmer, Karl E.
1971 Some general observations about nominal compounds. In Working Papers on
Language Universals 5: 1–24. Stanford: Stanford University Press. (Reprinted
in 1981 in Wortbildung, Leonard Lipka and Hartmut Günther (eds.), 233–257.
Darmstadt: Wissenschaftliche Buchgesellschaft Darmstadt.)