/1

On the General Semantics of Empty

I remember feeling perplexity every time when as a schoolgirl I was asked by my teacher to
“analyse” a sentence “syntactically” and “morphologically”, as it was then called. I followed the
given pattern of questions obediently - who, what, what is he doing, where, etc., and all was right
except when I came to strange little words about which no such questions could be asked, which
were vaguely and unsatisfactorily defined, and also, often neglected.
These, of course, were what we here designate as “empty” words (vs. “full” words). I am far
from considering this term satisfactory - it is just one of the variety of others, such as: secondary
words, auxiliaries, link words, discourse connectors, discourse items, logical signs (vs. descriptive
signs - Carnap 1965), etc. I hope it will be possible to have a fuller and a better name in the future.
Empty words are the words traditionally grouped into the following parts of speech: interjections,
particles, articles, prepositions, pronouns, conjunctions - not all of them are necessarily manifested
in every language, though. Some adverbs also belong to these, and occasionally, in text adjectives
and nouns of certain general semantics could function like “empty” words.
Our principal interest here lies in what the general functions of these “little” words are in the
language. We’ll first try to show briefly the most essential differences between full vs. empty words.
Full words have the following general qualities: they are semiotic signs, i.e. they denote things,
events, properties, etc., of the outer world. These signs are members of the system, which means
that the meaning and value of these discrete elements are determined by the relationship to the
other members within the system. Besides, full words are hierarchically built into complicated
morphological structures. They belong to paradigmatic classes and change their forms accordingly.
Consequently, in the process of communication full words perform two important functions - semantic
and grammatical: a. full words present the semantic material for building the lexical meaning and
the theme in the text; b. full words function as sentence members - they structure sentences
grammatically. Every full word can be substituted by another full word of the same paradigmatic
quality in the formal structure. E.g., the English word “daffodil” can have certain semantic functions
in a sentence (which is the most complicated structural unit itself) and be substituted by “rose”,
“chair”, “hope” etc. Owing to its denotative and connotative meaning the word “daffodil” also
introduces some kind of pastoral theme into a text. (Certain expectations are aroused in a listener
as to the occurrence of other lexical units in the sentence, and concerning the theme of the expected
linguistic expression).
Empty words are principally uninflected forms, they are not members of the formal structure,
they cannot be independent sentence members. The “substitution of the given element with another
one of the same class”, as it is formulated in the case of full words, is therefore, non-relevant for
empty words. Their main and essential function is of semantic character.
As we have argued elsewhere, empty words are close to indexical signs of Peirce’s terminology, i.e.
they are signals of some semantic notions and relationships. While a full word names certain logical
connections like, e.g., “addition”, an empty word “and” introduces it directly into the text. While a
certain “John” is named in the text with a proper noun, the pronoun “he” signals its re-occurrence
2/

in the text, and so on. Thus semiotically empty words are indeces either for something in the outer
world, or for something already mentioned in the text.
Comparing dictionary definitions for full words and empty words is striking, for the former
presents explanations, synonyms, visual material, but the latter - only verbal context. The meaning
of an empty word in a good dictionary is demonstrated with a deliberately chosen context - examples,
samples of texts, which provide a reader with the knowledge of the correct usage of the given empty
word. The meaning and the value of empty words are thus determined by their usage in the context.
According to Benveniste (1969), semiotic signs, i.e. full words, should be recognized, but semantic
units - units of speech, should be understood. Recognizing means being conscious of the identity
of the former and the latter usages, but understanding means perceiving the meaning of a new
semantic unit. A lexical unit of a language system - a semiotic sign, a full word - is principally new.
The justification of a system is in the possibility of transforming old meanings into new, descrete
elements into continuous units, the social quality - into the individual one, the neutrality - into the
subjectivity. (See also Weinrich 1976). This is the transition from recognizing to understanding.
This changing from semiotic signs and the language system to speech units (texts), according
to Benveniste 1969, is complicated, and it is difficult to know what subtle processes carry out the
transition. Now, to me empty words seem to be the means for the realization of these changes.
In every text there is a kind of “metasemantics”, says Benveniste, that is built over the lexical
meaning supplied by full words. Building this “metasemantics” is, I suppose, the most general and
linguistically most important function of empty words. Empty words ensure the transition from
formal hierarchical structuring and from the language system with its discrete elements, to a new
continuous communicative unit - a text, which is semantically structured in essence. The lexical
homogeneity provided by full words is perhaps only raw material for the further semantic structuring
of text. The essential features of the meaning of a communicative unit - its continuous and connected
character, completeness, and coherence to its situation are ensured by the work of empty words -
those “little”, “secondary”, “short words that help to economize”, as they are often qualified (de
Beaugrande and Dressler 1983).
The meanings of particular empty words in texts are being actively studied now. What we
attempt here is commenting on their general semantic functions.
Three major kinds of general semantic functions are ensured by empty words in building the
above-mentioned “metasemantics” in communication:
• building logical and semantic coherence,
• marking out new vs. old information,
• making the meaning cohere with the non-verbal context.
Structuring the text logically and semantically - the semantic function mentioned first - concerns
the inner semantics of the content of the continuous text. Here empty words signal the semantic
connections between text parts, like: causation, reason, result, inference, alternation, negative alter-
nation, contrast, addition, generalization, condition, reformulating, negation, concession, restriction,
numeration.
When we say empty words “structure” a text according to these semantic features, we mean
exactly that, for the parts of such semantic constructions could be substituted without causing any
destruction to the quality of the given semantic relationships. From this point of view the logical
and semantic connections inside the text are of the same quality as formal linguistic structures. A
simple example will suffice to illustrate the nature of semantic structuring carried out by empty
words:
(1) You can’t go - because - you are too young
You can’t do it - because - you are not strong
I’ll do it - because - I love him
 /3

Such semantic structures in texts are built by conjunctions, prepositions, particles, pronouns.
The next kind of major semantic functions of empty words is differentiating the old from new.
Saying something new about something old is really basic for human communication. Not only
grammatical patterning, deep structure relationships or word order, but also empty words contribute
to qualifying what is old and what is new in the given context. Two kinds of old and new information
could be distinguished here: one is the situationally determined new vs. old knowledge common
to the speaker and the listener. New information could in this case be indicated in the text by
particles universally. Particles can change the neutral old-vs-new patterning in the sentence by
placing emphasis on certain parts. The meaning of definiteness expressed through definite articles,
pronouns, indicate old information - something known to both participants of communication. The
notion of indefiniteness underlines new information, unknown to the listener and, possibly also, to
the speaker. Normally the speaker attempts to be carefully explicit about distinguishing what is
situationally old and what is new in his text. If this attempt is a failure, the communication does
not work.
Another kind of old vs. new information is determined by the verbal context - something has
been mentioned earlier in the text, something is repeated and known. Old information is shown by
co-referential devices built by pro-forms, and also by substitutional devices forming different kinds
of anaphoric and cataphoric ties, relying on the sameness of reference of e.g. full nouns on the one
hand and pronouns on the other; on the identity of a whole passage and a pronoun, etc. The same
is to be said of the pro-verbal forms. This is in fact the most usual and very active way of reminding
what is known to the interlocutors from the former samples of the text.
Another kind of co-reference - substitution - is employed when an empty words refers to different
elements from the identical class of words mentioned earlier. So the sameness is only partial, the
old repeated information is only partially identical. The device of substitution in English is often
formed by an indefinite pronouns “one” and “else” (e.g., “another one”, “a better one”, “someone
else”, etc.). Here is an example from Halliday and Hasan (1976):
(2) This is a fine hall you have here! I’ve never lectured in a finer one.
Substitutions like this are not equally usual to different languages. In Georgian, e.g., they are
comparatively rare - the corresponding device used in this language is ellipsis - the zero stage of
substitution.
The third major semantic quality of empty words - making the meaning coherent to the non-
verbal context - can also be of two kinds. One is the subjective character of the text that rests
upon speaker-oriented (also listener-oriented) modality. Attitudinal relations to the described events
could express: degree of belief, evaluation, ability or necessity, permission, possibility or impossibility,
hesitation, emphasis, wish, determination and the like. In Georgian, these are very actively expressed
by various particles, interjections. E.g., the particle kinagam expresses negative modality, but titkmis
implies positive attitude to the event (the meaning of both is “nearly, almost”).
Another situational characteristic is “spatio-temporal coordinates of the utterance”, as Lyons
(1977) puts it. Fillmore 1975 talks additionally of social deixis (the choice of honorific or intimate
speech levels), person deixis (the identity of interlocutors in communicative situation). Systems of
expressing deictic meanings could differ strongly in languages, which is well known and wil not be
discussed here. Deictic meanings are introduced into the text by personal pronouns and demonstra-
tives, prepositions, preverbs and postpositions.
Such seem to be the three kinds of general semantic functioning of empty words in text: Logical-
semantic coherence, marking out old vs. new information, building semantic ties with a non-verbal
context. Some words could be added about other semantic features of empty words
Empty words can be strong semantic dominants governing the occurrence of certain lexical
meanings and some grammatical features in the full words to which they are clitics. Georgian
titkmis, lamis, kinagam are clitics to verbs and mean incompleteness or non-occurrence of the action
4/

or event. They determine the aspect - the duration or momentariness, also lexical meanings in the
verbs to which they are attached; they also have positive, negative or neutral implications.
Empty words are always unambiguous in the context, expressing the one meaning that is de-
termined by it. In fact, the correct use of empty words is one way of ruling out ambiguity from
the text. Empty words never mean more than one thing at a time. Nonsense literature allows all
kinds of puns, neologisms, play on words with full, but not with empty words, just because of the
unambiguity of their usage.
The logical-semantic structuring ensured by empty words is what cannot be changed (or broken)
within one span of the text, and the constancy of this backbone is what makes the punning acceptable
and funny for the reader. English Nonsense Literature suggests plenty of examples for illustration.
Empty words are inherently polyfunctional, but in communication their functioning is quite
definite and unambiguous - no wonder that their quantity in a language is both finite and small,
but they work hard. There are always many empty words in any sample of text - their hard work
guarantees the full use of the exceptional ability of communicating in the way only humans can.
Only through empty words is the full expression of speaker’s intended meaning possible, they are
thus determining the quality, adequateness and appropriateness of the linguistic aspect of locution.

References
Benveniste, E. 1969. Semiologie de la langue:“Semiotica”. Mouton, La Haye.
Carnap, R. 1965. Foundations of Logic and Mathematics: The Structure of Language. In Readings in the
Philosophy of Language., ed. A. Fodor and J. Katz. Englewood Cliffs, New Jersey.
de Beaugrande, R., and W. Dressler. 1983. Introduction to Text Linguistics. Longman.
Fillmore, Charles. 1975. Santa Cruz Lectures on Deixis 1971. University of California Berkeley.
Halliday, M, and R Hasan. 1976. Cohesion in English. Longman.
Lyons, J. 1977. Semantics. London: Cambridge.
Weinrich, H. 1976. Sprache in Texten. Stuttgart: Klett.