File Information
File: 05-lr/acl_arc_1_sum/cleansed_text/xml_by_section/abstr/69/c69-2801_abstr.xml
Size: 43,344 bytes
Last Modified: 2025-10-06 13:45:44
<?xml version="1.0" standalone="yes"?> <Paper uid="C69-2801"> <Title>On the Use of IAn6uistiC/ Quantifying 0pera~ors in the Logico-Semanttc Structure Representation of Utterances</Title> <Section position="1" start_page="0" end_page="22" type="abstr"> <SectionTitle> I. Introduction </SectionTitle> <Paragraph position="0"> By now it has become olee~ to most linsuists that surface structures of utterances in natural laD6uage are not a sufficient basis for the semantic interpretation, an~ that an abstTact ~eep structure representation has to be assumed in a lin6uistlc desc~Iption, whlch la supposed to correlate the surface strua~xre hierarchical representation of each utterance with its msaniDg (or meanings in case of homonimi~y)deg Since the problem of what should be understood as the meanln6 of an utterance is not clear, the question of how deep etTuctures should be represented remains open for linguists, even for those who have been known as adherents of the transformational generative 5Tammar, where deep structures are determined in terms of generalized phrase markers (or some other similar concepts which get modified in the course of work).</Paragraph> <Paragraph position="1"> The questions which arise in postulating deep structure representations are often due to the lack of a possible guide llne C/oncernin6 the division between the knowledge of l~a~e and knowled6e of the world, and thus between the information which should be included in a full description of a lan6ua~e a~ that which shoul~ or need not. The classical distinction between syntax, which concerns relations between lan~ua6e -2signs, and semantics, uhich columns relations bergen lan~a6e signs and extraliD6uis%-lc &quot;objects&quot; (or extraliz~aistic relations), has also become blurred up by the fact of introducinK into the lin~uistic des~pttem abstract structures which are only indirectly related to the surface signs representin~ the elements t2~at actua/ly occur in utterances.</Paragraph> <Paragraph position="2"> However, even if we reckon the abstract deep structure deecrip~lons as pertainin6 to syntactical relations - in spite of cex~ain attempts tO the contreury 3- we will still be left with difficulties in trying to establish ~he scope of semantic relations, that is, the relations between linguistic signs which occur in deep structure rep~esent~a~ions and extralln-Kuistio &quot;objects&quot; or relations.</Paragraph> <Paragraph position="3"> The difficulties whi=h arise in establiahln~ the scope of semantics are before all due to the fact that the overhel~n 6 majority Of utterances contain siKns mhich are used in reference to the addresser, addressee, t~e time, place and situation in which a given utterance is produced2. Therefore, if we wanted to take s~tously the requizement that semantic relations should consist in relat~ S~DS with extTa~stic &quot;objects&quot;, such a task would s~ply be unfeasible for a lin~/st or semanticist, for it is cleeur that a~y two identical utterances used at a different time, place, or produced by a different speaker, should be inte:preted differently if they contain &quot;token-reflexive&quot; signs, ~ obviously they are related to two different extr~stic situations.</Paragraph> <Paragraph position="4"> -3-It is certain~ much more reasonable to follow a view held by some philosophers, logicians and linguists, which can be briefly recalled by means of a quotation from Roman Jakobson: &quot;the meanln~ of any linguistic sign is its translation into some further, alternative sign, especially a si@u &quot;in mhich it is more fully developed&quot; as Peirce, the deepest inquirer into the essence of signs, insistently stated &quot;3. The general idea is certainly right, but such an approach leaves much freedom with respect to various possible ways of presentin~ &quot;the alternative signs&quot;. Ne will not discuss here the different proposals connected with such an approach to semantics, their assets and drawbacks, but - accepting in principle the thesis that a semantic model for natural language can only be established by means of &quot;alternative signs&quot; - we will suggest one of the possible ways of describi~ the semantic interpretation of utterances.</Paragraph> <Paragraph position="5"> Following a concept which has already a certain tradition in philosophy, semantics an~ logic ~, but has never been taken seriously by linguists, I will accept the view that the semantic interpretation c~ an utterance is the set of consequences which can be derived from that utterance (or more strictly, from its deep structure description) on the basis of certain implicatienal rules that can be established for the given language. Accordingly, I attempt to search for such a description of deep structures for which it ,/ill be possible to formulate the rules ~hich i call quasi implications 5. In order to avoid confusion in using the term &quot;deep structures&quot; (which by now is employed in different senses, but should be reserved for transformational deep structures, for which it has been clearly defined), I will refer to the logico-semantic structure (LS structure) of utterances 6.</Paragraph> <Paragraph position="6"> The identification of the semantic interpretation of an utterance with the set of consequences or conclusions which can be drawn on the grounds of that utterance is very close to our intuitive understanding of how we interpret utterances.</Paragraph> <Paragraph position="7"> To say that the addressee understands an utterancep only if he is able to draw all the conclusions implied by the given utterancep seems to be compatible with the ordinary thinki~ and the estimation of a fully adequate understanding of an utterance. The fact that certain conclusions are based on premisses corresponding to the addressee's beliefs which belong to his knowledge of the world precisely reflects what has been mentioned earlier about the lack of a clearcut division between the knowledge of language and knowledge of the world.</Paragraph> <Paragraph position="8"> In fact such a division can only be arbitrarily established for a given description of language. For instance, in describing a lexicon (or the rules of the lexicon in a theory of language), we have to make a decision as to Whether certain information concerning a lexical item should be included in the lexicon~ or treated as encyclopaedic information, which pertains to such or other domain of science, specialization, etc. The arbitrariness of this division ooncerns, l however, usually the boundary problems with respect to which such questions arise, but it does not effect the fact -5that in the majority of problems there is a general agreement of lin6uists as to what syntactical and lexical information should be included in the linguistic description in some or other way. Once such a decision is made, and the information is included in a lin6uistic description, we should be able to define clearly what consequences can be drawn on the basis of a given utterance and linguistic quasi iMplications, and what ones require additional premisses which belon~ to the knowledge of the world.</Paragraph> <Paragraph position="9"> To give an example, we may say that a speaker or author uses appropriately 7 the utterance : (1) ~ixon has left the United States only if his purported belief is that : (la) there is one and only &quot;object&quot; which is bein~ referred to bj him as &quot;Nixon&quot; (Ib) there is one and only &quot;object&quot; which is being referred to by him as &quot;the United States&quot; (1C/) Nixon was in the United States before and if his purported claim is that : (1~) Nixon is not in the United States at that time.</Paragraph> <Paragraph position="10"> The propositions (la, lb, lc and ld) can be derived formally from (I) on the basis of certain linguistic quasi i~plications 9, and they below, among ethers, to the set of consequences which constitutes the interpretation of (I). On the other hand, the proposition (le) The President of the United States has left the United State s is a conclusion based on the premiss :&quot;Nixon is the President Of the United States&quot;~ which may belong to the speaker's and heater's knowledge of the wOrld, but if not included in the -6description of the lexicon, it will not be derived formally as a consequence.</Paragraph> <Paragraph position="11"> It is worth noting that the propositions (la, Ib, ic) concern the speaker's purported beliefs, whereas the propOsition (ld) concerns the speaker's purported claim or assertion. The difference is significant, for if we convert (I) into a ~aestion, negation, wish, command, etc (Has Nixom left the United S~ates. E ~ixon has not left the United States, I want Nixon to leave the United States, etc) then the consequences pertaini~6 to the speaker's beliefs will remain the same, and it is only the consequence (Id) pertalnin6 to the speaker's assertion that will become oorrespomdln61y d/fferent. In general all the consequences derived from an utterance can be accepted only as a set of propositions which correspond to the purported attitude of the speaker, n.~ as a set of propositions which hold true. The question of w~ether such propositions are true or not is not a linguistic question, similarly as the problem of whether the &quot;object&quot; being referred to exists in reality or is to be assumed in the context of a novel 10 In terms of the approach suggested here~ontological conslderations have no beariD6 on the LS structure description of utterances, from which ~he same consequences can be d~awn~independently of whether the speaker is sayIn6 the l truth or lyin6, whether he is relati~ his dream, tellln6 a story or repcrtiDg facts.</Paragraph> <Paragraph position="12"> In this paper we will discuss only those aspects of the LS structure representation which concern certain liz~ist-?- null ic quantifyiD~ operator2. As in general we assume that the LS s~ructure could be represented in a language somehow analogous to that of a modified predicate calculus, our aim in the present discussion is to suggest a certain w~y of defining the use of linguistic quantifyi~, operators which would be in agreement with the interpretation of utterances in natural language ~ the present paper is concerned with 11 English utterances only.</Paragraph> <Paragraph position="13"> The problem of defining the use of linguistic quantifying operators in the LS structure description seems to be one of great significance for the semantic interpretation of the surface syntactic relations, for it appears that it is hardly possible to assign an adequate semantic interpretation 12 to an utterance, if its LS structure is not conceived of in terms of at least one predlcate and its ars~ment (or arguments) prefixed by some linguistic quantlfyi~ operator.</Paragraph> <Paragraph position="14"> AD~thing we talk about~ that is~ any subject (or sub~ects) of predication, is either a particular &quot;object&quot;, a particular relation between &quot;objects&quot; of some sort, a particular situatlon, event, stc~ or else it m~y be a certain number 13 of ~'objects&quot;p situations, etc which are such and such, or finally the subject of predication may be each &quot;obJect&quot;~ situation, ere, which is such and such. Generally speaki~, en utterance is analogous to a logical sentence or proposltion~ rather than to a propositional function with free variables~ and ~hus the LS structure of an utterance may be best conceive~ as a fo~nula in which ~e make use of some sort of lin@~ulst~ Io quantlfyin~ operators.</Paragraph> <Paragraph position="15"> -8-The analogy with the language of predicate calculus is not very precise though, and - as we shall try to show - the linguistic quantifying operators must be defined differently than these defined in logic. In the first place, there is a greater number of linguistic devices which function as quantifiers of some sort, and in the second place, even those devices which show an analogy with the operators used in logic have a slightly different semantic function and cannot be defined for our use in the same way as those defined in logic. In the present paper we shall discuss only two linguistic quantifying operators I@, one being analogous to the iota operator defined for the predicate calculus, the other one being analogous to the general quantifier.</Paragraph> <Paragraph position="16"> 2- The linguistic iota operator The iota operator, as defined in logic 15 can he used in the expression (~x)~(x),only if the propositional function is a unit function, that is, if it satisfies two conditions, one concerning the existence, the other the uniqueness of the &quot;object&quot; which is ~ . Such conditions ms2 be expressed formally as The expression (~x)~(x) has the category of an argument (not of a statement), and may be interpreted as the English expression of the category of a nominal phrase, namely, :&quot;that one and only object which is ~ &quot;.</Paragraph> <Paragraph position="17"> Now when comparing the expressions in English which have -9a similar semantic function as the unit function, namely, lingulstlc indices 16, it is immediately obvious that we are not in a position to prove that they satisfy the conditions of existence and uniqueness, on the contrary, it would be hard if not impossible to find a single lin6ulst~c index which, as such, would satisfy such conditions.</Paragraph> <Paragraph position="18"> ~evertheless, it is an empirical fact quite evlden~ ~hat whenever an addresser uses an utterance containiDg an index ~ in the normal ~ process of communicatlon, it is always in accordance with his purported belief that there is one and only &quot;object&quot; which is being referred to as ~ * In contradistinction to a formal system for which we have ~e define the rules and conditions of correct use, the linguist works in the opposite direction. For we alrea~7 have a given system which we all use successfully for the purpose of intercommunAcation, and the linguist's or the semanticist's job is to detect the rules that make possible the achievement of mutual understandln 6 among ~he competent users of the given language system* Thus in the first place we can analyse utterances only with the assumption that they are used appropriately - otherwise no consistent analysis would have been possible.</Paragraph> <Paragraph position="19"> The appropriate use of language, as defined for our purpOSeSD is nothing mere than a use which is consistent with ~he rules of language and with the speaker's purported propositional attitudes. It would be thus unreasonable ~o reject such an obvious assumption.</Paragraph> <Paragraph position="20"> -lO-We may trytthen)to establish quasi implications ~lich will held truelTand thus yield the correspondi~ consequences for all eases of lin6uistic use in the process of communication. Accordin61y, we will show that what is defined as a condition of correct use in a formal system, will be derivable in our oase as a consequence of a given utterance.</Paragraph> <Paragraph position="21"> In the second place, we take into account the fact that utterances of natural lan6uage are never, so to speak, hun6 in the air, but are always interpretable in terms of what we 18 call the propositional attitude p or modal frame 19 when referriD6 to its explicit representation. For no utterance used in the process of communication can be fully interpreted without our understandiD6 it as a proposition which expresses, amoD6 other things, the speaker's or author's propositional attitudes, which may be beliefs, assertions, doubts, requests (for oral or other responses) or any other attitudes.</Paragraph> <Paragraph position="22"> Accordin6 to such an argumentation, the propositions which correspond to the two conditions for the correct use of a unit function, (a) and (b), will be accepted in our case as consequences which beloD6 to the interpretation of any utterance ~ containin8 a linguistic index . Such propositions cannot,however, be represented as ocourrin~ in an assertorio modal frame, but in the modal frame :&quot;The speaker believes that ... &quot; (where the term &quot;believe&quot; is used in the sense of purported belief). The quasi implicational scheme on which such consequences are drawn is roughly speakin6 as follows: For any linguistic index ~ , the speaker uses appropriate-W ~ only if he believes that there is an &quot;object&quot; which ly -llis ~ , and that the &quot;object&quot; being referred to as ~ is uni que.</Paragraph> <Paragraph position="23"> Since the consequent of this statement can be assumed to express one of the necessary conditions for the antecedent 20, we ma~ consider the whole statement as a quasi iMplicatisn.</Paragraph> <Paragraph position="24"> Thus in the case of any utterance which contains a lin@uistaccept null ic index, we may ~ a statement correspon~ing to the antecedent as hold~ true - by virtue of our general assumption of appropriate use - end therefore we can always infer the corresponding consequent as a consequence.</Paragraph> <Paragraph position="25"> We thus preserve the uniqueness of the purported &quot;object&quot; by means of inserting into the common (and thus intersubJeetire) semantic interpretation of indices a parameter which isrttoken-reflexiverPand which serves its purpose(a~ time an utterance is actually used~ in relating signs to extralin~stic &quot;objects&quot; or relations,throug~h the identification of the speaker, the time, place and situation in which ~he given utterance is used. Accordingly;the value of the argument - which in the case under discussion is a linguistic index~ is understood in any particular case, without being possibly indicated in a general linguistic description. In refers other words~ the fact that the speaker ~ to one am~ only &quot;object&quot;! each time he is using an imde~ belongs to the semantic Interpretatlonpan~ is reflected in the propose~ LS structure representation and the corresponding consequences &quot;which ere derivable from It, but the question of what &quot;object&quot; me~ actually be identified by means of such an expression does not belong to the domain of lingaisticsdeg - 12 A~ such &quot;object&quot; m~y only be described in terms of linguistic expressions - it may either be pointed to by means of a name used as an index, or described in terms of the lexical items contained in the definite description used as an index. A linguist solves thus the problem of how an &quot;object&quot; is described or pointed to in order to be identified, and he ~ not be concerned with Iha___~t it is.</Paragraph> <Paragraph position="26"> It thus appears that the lin6uistic iota operator is a close analog to the iota operator as defined in logic, the dif~erences lying in our treating the conditions of correct use defined in logic as the consequences of the ever-assumed appropriate use of utterances, and in our introducing the propositional attitudes into the LS structure description.</Paragraph> <Paragraph position="27"> As it will be argued below, the next linguistic quantifyin~ operator to be discussed in this paper, namely, the all-operator, will also be defined partly by analogy to the iota operator, for the reasons based on linguistic evidence.</Paragraph> <Paragraph position="28"> The all-operator In the preceding section we have discussed a certain type of ars~Ament prefixed by the iota operator. Such an argument will be referred to as an argument of type I. Its explicit exponent on the surface can only be a linguistic index, and its role is to indicate one and only &quot;object&quot; to which a given predicate applies. At present we will discuss another type of argument - which will be referred to as an argument of type 2 - whose role is to indicate all the &quot;objects&quot; to which a given predicate applies. The explicit exponents on the surface of English utterances are usually nominal phrases preceded by &quot;all&quot;, &quot;every&quot;, &quot;each&quot;. -13-However, it is not always the case that such nominal phrases are to be interpreted as arguments of type 2 (consider phrases such as, for instance, &quot;all the country&quot;, every second French man&quot;, &quot;each time&quot;, etc). Let us now discuss the followln6 examples by means of which we will show the analogy between arguments of type 1 and those of type 2.</Paragraph> <Paragraph position="29"> (2) John has awaken (3) JY father has awaken (4) Ky parents have awaEen (5) All the members of my family have awaken (6) All the inhabitants of this house have awaken (7) All the inhabitants of the moon have awaken Independently of other problems connected with the proper LS structure representation of utterances in which a predi~ applies to each individual of a set determined by a given ~es- null cription, we have to admit that all the nominal phrases in our examples have a similar function ; they serve to identify either a single individual (utterances 2 and 2), or a certain number of individuals (utterances 4,5, 6 and 7), and thus serve to indicate the values of the argument used with the predicate &quot;has awaken&quot;. Thus if we compare : (8) The boy standing in the doorway is tall with (9) All the boys standing in the ~oorw~7 are tall it seems clear that the difference between the t~ lies in that the nominal phrase in (8) serves to indicate or identify one indlvldual~ whereas the nominal phrase in (9) serves to indicate or identify a certain number of individuals, both n@~inal phrases beinK used as &quot;instructions&quot; for identification, and as such correspond to the argument of the predicate &quot;tall&quot;. Such &quot;instructions &quot; are used approprlatelyponly if the speaker believes that there is a boy standing in the doorway (for 8 ) and that there is more than one boy standin 6 in the doorway (for 9). Consequently by virtue of our Keneral assmaption of appropriate use, it is possible to infer such pseudoexistential propositions in the model frame &quot;The speaker believes that * *..&quot; One might object to this interpretation of all-statements and try, by analogy to logical statements with a general quantifier, to interpret them as implicational statements :For a~ if x is ~ , then x is ~ . Under such an interpreta- xp tion the utterance (9) would be a paraphrase of -(i0) If anythlng is a boy standing in the doorway it is tall If anybody is a boy stsnding in the doorway he is tall Anybody who is a boy standin6 in the doorway is tall Even if we ignore ~he awkwardness of utterances such as (i0), we cannot accept such utterances as paraphrases of (9) for reasons. First, (i0) could, then, equally well be taken as a paraphrase of (9) as that of (87, as there is no uention with respect to the plurality of the individuals being referred to. Thus (i0) could be accepted as one of the consequences from (9), (and this will be shown below to be the case), but not as an equivalent proposition. Second,, the speaker implies by (9) his purportecl belief that there are individuals who are boys standing in the doorw~7, whereas no such izplication bholds for (10). To give a clearer example, if someone says (11) All the inhabitants of the moon must feel cold the heater's reaction might be expressed by ~'&quot;~at are ~ou - 15talklng about. ~- There are no inhabitants on the moon.&quot; On the other hand, no such reaction would follow if someone says : (12) If anything is an inhabitant of the moon it must feel cold The response might then be : &quot;Right you are, but fortunately there are no inhabitants on the moon&quot;.</Paragraph> <Paragraph position="30"> PSnother argument supportin6 this interpretation may be provided by the followin6 test. If we consider the utterance: (13) I doubt whether there are inhabitants on the moon, but all inhabitants of the moon must feel cold it leaves the impression of a certain awkwardness, ~hich is a result of the inconsistency between the expressed doubt and the i~lied belief concerning the same proposition. On the other hand, the utterance : (1~) I doubt whether there are inhabitants on the moon, but if anything is an inhabitant oZ the moon it must feel cold does not seem awkward, since according to our argumentation implicational statements do not imply ext7 belief or assumption concernlng existence. Evidently there are different i~lioational rules concerning the if.oothen-statements, and they depend on the tense used in the If-clauseo If we denote the if-clause by S~ the hea~er may infer, for instance~ that the speaker believes that S~ or believes that it is not the case that S~ or believes that it is probable that S~ etco In any case~ it seems clear for the reasons ~iven above that (I0) is not a paraphrase of (9), nor (12) is a paraphrase of (11). In spite of the fact that implicational statements can be considered as truthfunctionally equivalent to the corresponding all~statements (and this is the basis of the log~Ical equivalence)~ yet they cannot be represented by the same - 16-LS structure, for obviously the latter convey additional semantic information, that is, additional consequences can be inferred from them.</Paragraph> <Paragraph position="31"> AC/cordin61y, instead of making use of a general quantifier, we will introduce a quantifying operator which we mall all-operator. The all-operator is assumed to be an argument-forming eperator, not a statement-forming speratorp as in the case of a general quantifier in legic. In other words, it is not an operator which makes a statement when prefixed to a propositional form, but one which makes an argument when prefixed to a function representing a nominal phrase in the plural role form. It ~hus has the same ~ as the iota operator, which is used to make an expression of ~he category of an argument 23. We assume here that the all-operator belongs to a speclfic class of modifiers which we call ~ linguistic quantllying operators. Such operators will occur in the LS structure representation;always being prefixed to a variable x.</Paragraph> <Paragraph position="32"> A logical statement which is prefixed by a general quantifier with limited range * is an abbreviation for the i~lieational statement We will, however, use a different expression, provisionally denoted by (All x) ~ (x), which will not be interpreted as a statement equivalent to an impli'cati@nal statement, but - by analogy to the expression bound by the iota operator it will be interpreted as an expression of the argument type : all the &quot;objects&quot; which are ~ . According to our interpretation, the corresponding implicational statement - 17 will be accepted as one of the consequences which follow from all-statements.</Paragraph> <Paragraph position="33"> By analogy to the case of the iota operator, we can here also establish a quasi implication concerning the use of the all-operatorpon the basis of ~he empirically given conditions of appropriate use of all-statements Consider the following propositional schemes: (a) There is more than one &quot;object&quot; which is ,or any &quot;objeot-. if it i, . it is The two propositional schemes correspond to propositions which may be considered only in terms of certain proposi~ionsu% attituPSes of the speaker who is ut-beriag an allstatement. We ma~ define a general quasi implicationsul scaeme whiCh, roughly, will be as follows: A speaker uses appropriately an al ! statementponly if his purpoz~ed belief is that there is more than one &quot;object&quot; w~ich is ~ , and his pu~orted claim is that any &quot;object&quot; which is ~ is ~ * Now by our general assumption that we are concerned only with those utterances which are appropriately used, we can always derive a proposition corresponding to (a) as representing the speaker's belief, and a proposition correspondtng to (b) as the speaker's claim or assertion, and both belong to the set of consequences of an all-statement.</Paragraph> <Paragraph position="34"> The difference between the speaker's propositional atti%-ude with z~spect to (a) and (b) is analogous ~hen ~e convert an all-sta~ement into a corresponding negation, question,</Paragraph> <Paragraph position="36"> request, doubt, etc, for it will appear that the proposition ~orrespoD~ling to (a) in the~ frame &quot;The speaker believes that ...&quot; can be derived as a consequence, whereas the proposition oorrespondi~ to (b)lalthoug~a also derivable as a consequence, will occur in each ease in a dlfPSerent modal frame, ~hich will express correspondingly the various propositional attitudes of the speaker (The speaker denies that deg.., q~ae speaker wants to know if .deg., etc) For instance, for the utterance which is a ~uestion : (17) ~e all the inhabitants of this house sick. ~ we have a set of consequences to which belong, among other propositions, the following two : (18) The speaker believes that there is more than one &quot;object&quot; whlch is an inhabitant of this house (19) The speaker wants to know if eu~ &quot;object&quot; which is an inllabitant of this house is sick.</Paragraph> <Paragraph position="37"> These aonsequsnces are obtained on the basis of a very ~eneral quasi impllcational scheme conoerniD~ all-stateRents converted into this type of questions.</Paragraph> <Paragraph position="38"> In conclusion it may be ~rth emphasizin~ that by intro* uciDg propositional attitudes into the lin~Aistic desc~il>tion,a~i by defining quasi i~plications in terms of propositional attitudes, we are able to derive the correspondlng consequences not only from declarative sentences, but from all ~y~es of utterances. Thus the scope of our rules of inl null ference ~ is much wider than the scope of such rules in logic. Accordln~ly the identification of the semantic in~ terpretation of an utterance with the set of its consequences is possible for all types of utterances. 2~ - 19 Footnotes null 1. See U. Wein~eich, Exploration in Semantic Theory, in Current Trends in Linguistics, Vol. 3, e4. T.A. Sebeok, Mouton, The Ha6ue, 1966, where the author argues in favor of an approach which makes no attempt &quot;to fence off mutually exclusive domains for syntax and semantics&quot; and thus he is against the assumption that semantics begins where syntax ends. Deep structures are conceived of by him in terms of both the syntactical and the semantic relations.</Paragraph> <Paragraph position="39"> 2. Such signs have been widely discussed by logicians and llnguiets. They are referred to as token reflexive signs (by Hans Reichenbach), as egocentric particulars (by Bertrand Russell), and as shifters (by Roman Jakobson~. null 3- Roman Jakobson, On Linguistic Aspects of Translation, in On Translation, ed. R.A. Brower, Oxford University Press, New York, 1966, p. 232.</Paragraph> <Paragraph position="40"> @. See C.J. Lewis, The Models of HeaDing, Philosophy and Phenomenological Research, Vol. IV, No 2, 19~, for his conoept of connotation or intention, which bears an in~an~l G. Frege'~ fluence of J.S. Uill~s * ee a so R. Carnap, Empiricimm, semantics and ontology, in Revue Internationale de Philosophie~ No @, 1950, where the author accepts the view that the set of consequences following from a senfence may be identified with the meaning of that sentence. Such an idea was also professed by H. Hi~ during his lectures on semantics (University of Pennsylvania, - 20 * 1965).</Paragraph> <Paragraph position="41"> 5- The conoept of quasi implications was Introduced in my papers : On a Condition of %he Coherence of Texts, paper delivered at the International ~mposiue on Semiotics, Warsaw, Aueust, 1968 j On the Semantic Interpretation of Subject-Predicate Relations in Sentences of Particular Reference, to appea~ in Progress in Lln~tistics, eds. Bie~rlsch and HeidolI~h, Mouton et Go., ~he Hague ~ Ar~mente and Predicates in the Logico8emanti@ Structure of Utterances, to appear in Studies in Syntax and Semantics, ed. F. Kiefer, Foundations of Language Supplementamy Series, D. Reidel et Go., Dordrecht-H elland.</Paragraph> <Paragraph position="42"> 6. The term logico-semantic structure was used in my foamer papers (see footnote 5)7- In 5eneral I say that a speaker uses appropriately an expression or an utterance, if he uses it consistently with the rules of languaKe an~ with hie purported beliefs (see footnote 8, for the term &quot;purported belief&quot;). For a consistent semantic interpretation of utterances, we have to consi4er only those utterances which are used appropriately, as the appropriate use in this sense of the term is nothin~ more than the assumption ~hat the speaker knows the lan6uage he is using and says only what he intends to gay ( he makes a correct use of an utterance accor~i~ 8 with what he wants to say). 8. From the point of view of the semantic interpretation, it is irrelevant whether the speaker in fact believes - 21 or only pretends to believe that such and such is the case. I therefore use the term &quot;purported belief&quot;, and ~nenever I use here the expression &quot;The speaker believes that ...&quot; it is only in this sense of the term. The same holds true of any other propositional attitudes.</Paragraph> <Paragraph position="43"> 9- Linguistic quasi implications can be defined for the use of lexical items and for the use of certain structures. In our example, the consequences (la) and (Ib) ere based on a quasi implication concerning the use jof linguistic indices (see Section 2, below), whereas the consequences (Ic) and (Id) are based on a quasi implication defined for the lexical item &quot;to leave&quot; (For more details, see the section on implicative terms in my paper On a Condition of the Coherence of Texts, op. cir.) 10. The term &quot;object&quot; is used here also in the sense of &quot;purported object&quot;, that is, for anything namable. Evidently the knowledge of the type of discourse (everyday, scientific~ literary, etc) may tell us whether the speaker or author believes that a given &quot;object&quot; exists in reality, or it is to be assumed as in a hypothetieal theory, or imagined as in a novel. But such n considerations have no bearing on our present discussion.</Paragraph> <Paragraph position="44"> 11. In general, it seems that the concepts we are discussing concern other languages as well. If in a given language there are no explicit exponents on the surface which would correspond to the interpretation we are assuming, for instance, for linguistic in- 22 dices used in English, the~e are probably some way to distinguish the utterances in terms of similar conaepts, for it is hard to imagine how people could manage to communicate without maklr~ use of this rather general concept and others discussed here.</Paragraph> <Paragraph position="45"> 12. By an adequate semantic interpretation I mean such an interpretation which is compatible with the lin~alstic intuitions of coherent speakers, and testable a~ainst the practice of linguistic communication. Examples of * such semantic tests are given below (p.14,15), 13. The number may be exactly specified by a numeral, or expressed in an indefinite way, for instance, by words such as &quot;man~&quot;, &quot;few&quot;, &quot;the majority of&quot;, etc. 14. See my paper, Arguments and Predicates...&quot;, op. clt. for a rough discussion of Other quantifying operators.</Paragraph> <Paragraph position="46"> 15. The iota operator was first defined and used by Peano, then by Russell, Reichenbach, Mostowski and several other logicians.</Paragraph> <Paragraph position="47"> 16. A lin~ulstio index was defined in my paper, On the Semantic Interpretation .. :, op. cir. Rou6hly, a liD6uist-ic index is a personal pronoun, a proper name or a definite description (that is, a nominal phrase preceded by a definite determiner, used in an identifying role) 17. In general the quasi implications which we define hold true by virtue of the empirically ~iven conditions.</Paragraph> <Paragraph position="48"> The truth of the consequent in our quasi ilpllcations always constitutes the necessary condition for the truth -2~of the antecedent, and thus we may consider the mhole as a valid conditional (material implication).</Paragraph> <Paragraph position="49"> It may be interesting to compare the various approaches to the concept of&quot;presupposltions&quot; (as referred to in the literatu~re),which we describe here as consequences. (See G. Fre~e, Sinn und Bedeutung, Zeitschrift fur Philosophie und philosophische Kritik, 1892 ~ P.F. Strawson, On Referring, ~ind, 1950; O. Ducrot, Los pr~supposSs, conditions d'emplo~ ou @l@ments de contenu, paper delivered at the International Symposium on Semiotics, Warsaw, 1968 ~ A. Wierzbicka, 0 sp6jno~ci semantycznej tekst-u wielozdaniowego, paper delivered at the Conference on Semiotics, IBL, Warsaw, 1968 ~ C.J. Fillmore, Types of Lexical Information, Working Papers in Linguistics, i~o 2, The Ohio State University, november 1968~ * In the present approach, presuppositions are identified with those propositions which belong to the set of consequences following from a given utterance and are provided with the modal frame :&quot;The speaker believes that...': Such propositions belong accordingly to the semantic interpretation of a given utterance. This approach is close to that of O. Ducrot (Les presupposes..., op.cit.), who also treats such propositions as constituting part of the meaning of a given utterance, rather than as the required conditions of use9 or as propositions which are presupposed to be known by the hearer, etc. The difference beSween the two approaches (which have been developed independently of each other) consists thus in my attempt to -24account for this element of the mani~ of a Given utterance by scans of derivin~ such propositions as consequences baaed on the corresponding quasi i~lications defined for partlcule~ lexical items and particular structures. Such a treatment is consistent with the general proposal to identify ~he meani~ ~ an utter~ ance with the set of its consequences.</Paragraph> <Paragraph position="50"> Fillmore, on the other hand, distinguishes between &quot;the presupposltional aspect of the semantic structure of a predicate and the &quot;meenln~&quot; proper of the predicate&quot;. He identifies the presuppositions of a sentence as those conditions which must be satisfied before the sentence can be used in an~ of its functions. Thus the sentence &quot;Please open the door&quot; according to Fillmore (Types of Lexlcal luformation, op. cit)&quot;cen be used as a command only if the TL &quot; (that is, the addressee) &quot;is in a position to know what door has been mentioned a~l onl~ if that door is not at TLA&quot; (that is, the time of producing the utterance) &quot;open&quot;. According to my approach, such an utterance will be interpreted in any case as a command, for its interpretation cannot be dependent on whether the above mentioned conditions are,or are notpsatiefied. Suppose I hear somebody ringing the bell add I ask my sister Please open the door&quot;. I~ may happen that the ~oor is open at that t~me, but the state of the door (its bei~ open or closed) has no bea~i~ on the interpretation of this utterance - which in any case is intende& by the addresser, and interpreted by the addressee, as a command. Thus, as it appears from the evidence of fan.age use, it is only the addresser's purported belief :&quot;The door is not open&quot; which may be safely assumed with respect to the given utterance.</Paragraph> <Paragraph position="51"> The term &quot;propositional attitude&quot; is adopted from Bertrand Russell, Inquiry into Meaning and Truth (Penguin Books, Baltimore-Maryland, 1962).</Paragraph> <Paragraph position="52"> The term &quot;modal frame&quot; has been coined by A. Wierzbicka (0 sp6jno~ci semantycznej..., op. cir, and her o~her papers). I use the term &quot;propositional attitud~'when referring to the content oi&quot; an utterance, and the term &quot;modal frame&quot;, when referring to its explicit representation (which in the present paper is only rendered in words ).</Paragraph> <Paragraph position="53"> This thesis is based on empirical evidence. If we seem to find apparent counterexamples, it always appears that they are cases of linguistic misuse, and thus they do not falsify lu~ the quasi implication, in such oases the antecedent is false. We are concerned, of course, only with the normal use of utteranoes in the process of communication, not with cases in which utterances are used as examples, when indices are not supposed to identify anything.</Paragraph> <Paragraph position="54"> There remains one more remark to be made in connection ~ith our quasi implicational scheme. The copula &quot;is&quot; ~aich is used in the consequent should be interpreted as, so to speak, tenseless, for its tense is dependent - 26 the tense of the verb in the predicate which applies to the &quot;object&quot; referred to by ~ in a given utter ~ a~lce.</Paragraph> <Paragraph position="55"> 21. The definite description used as an index is best conceived of as an &quot;instruction&quot; for identlfieaticn. The linguist is thus concerned with what the &quot;instruction says, rather than with the problem of what a given &quot;object&quot; is. The &quot;object&quot; in our sense of the term mej be conceived of as just the value of the argument * 22. Notice incidentally that there is a differenae between utterances such as (4) and, for instance, &quot;XY parents bought a house&quot;, where the desorlp~on &quot;my parents&quot; should be represented as a linguistic index (argument of type I~ which has a unique value) by means of which we identify one and only neglect&quot; (a couple of individuals as a ~hole ) to which the ~iven predicate applies. The predicate is not used here distributively as in the case of (~), inhere it applies to each individual referred to by the description &quot;my parents&quot;.</Paragraph> <Paragraph position="56"> 23. The expression (~ x)~(X) is interprete~ as &quot;that one and only object which is ~ &quot;, add not as &quot;there is one and only object which is V &quot;.</Paragraph> <Paragraph position="57"> 1 24. Finally, I wish to excuse myse f for a ratheE loose way I am using certain terms which are well defined in lo~ic (for instance, arE~,ment, value). I hope , \ - 27 however, ~at the reader will get from my rough presentation the underlying linguistic concepts which I a~temp~ed to submit for discussion.</Paragraph> </Section> class="xml-element"></Paper>