Semantics of negative pronouns in the light of compatibility with the modifier "almost"

Rossyaikin Petr Olegovich ORCID: 0000-0002-7235-724X Senior Laboratory Assistant, Faculty of Philology, Department of Theoretical and Applied Linguistics, Lomonosov Moscow State University 119991, Russia, Moskva, g. Moscow, ul. Leninskie Gory, 1, str. 51 petrrossyaykin@gmail.com

Introduction

The languages of the world can be divided into two groups – languages with double negation and languages with negative agreement. In languages with double negation, such as English, negative pronouns (nobody, nothing, etc.) can express negation independently (1a). If there is a verbal negation in the sentence, then it is interpreted with a double negation (1b). That is, one negation is introduced by a pronoun, and the second by a verbal negation.

(1a) Nobody came. ‘No one came.’

(1b) John didn't see nobody. ‘It's not true that John didn't see anyone’ = ‘John saw someone.’ (double negation)

The Russian language (like other Slavic languages and some unrelated to them, for example, Hungarian ^[1]) refers to languages with negative agreement. In these languages, negative pronouns cannot express negation on their own (2a). Only sentences with a double (triple, etc.) negation expression are allowed – as part of a pronoun and with a verb (2b). (Below we use * to denote grammatically incorrect constructions and sentences.) In this case, the negation is semantically interpreted only once. In particular, the academic grammar of the Russian language ^[2] offers the following generalization about such sentences: "In a sentence with a negation, pronominal words with a prefix. ni, word forms with the particle ni and rows with the conjunction ni...neither they do not perform an independent function of negation: they serve only to strengthen negation, accentuate it, give it an exclusive character" ^{[2, p. 412]}.

(2a) *No one came.

(2b) Petya didn't see anyone. ‘It's not true that Petya saw anyone.’ (single negation)

The fact that negation is expressed twice (or even more times, as in a sentenceNo one has helped anyone in any way), but is interpreted only once, presents a well-known problem in the analysis of the semantics of negative pronouns in languages with negative agreement, to which a large number of works are devoted, as in the Russian ^{[3],[4],[5],[6],[7],[8],[9]}, and in foreign linguistics ^{[1],[10],[11],[12],[13],[14]}. However, there is currently no consensus on this issue.

The purpose of this work is to establish the semantics of negative pronouns in languages with negative agreement on the material of the Russian language. In order to establish the semantics of the language unit under consideration, it is necessary to analyze the possibility of its use and interpretation in various contexts. This article is devoted to the consideration of constructions of the type almost nobody.

The fact is that the modifier can almost not be combined with all quantifier words and pronouns. For example, the sentence almost every student solved this problem is permissible, and the sentences *almost some student solved this problem or *almost anyone solved this problem are unacceptable. The permissibility or inadmissibility of certain pronouns and quantifier words in combination with a modifier is almost due to the semantic or syntactic properties of these lexemes. Thus, the relevant data can be used as material for semantic and syntactic analysis.

In this work, we use the examples we have constructed as a data source. Most of them are elementary and serve exclusively illustrative purposes. They are unequivocally evaluated by native speakers as grammatically correct or incorrect. The main interest is the examples discussed in section 4, which, as far as we know, have not been considered by previous researchers. The acceptability assessments of these examples are based on a survey of native Russian speakers.

The article is structured as follows. In section 1, we consider the variants of semantic analysis of negative pronouns in the Russian language. In section 2, we consider the semantics of the modifier almost. Section 3 is devoted to the discussion of one of the variants of the analysis of combinations of the type almost nobody, which we will refute in the future. In section 4, we present new data and analyze the semantics of negative pronouns in the light of these data. Section 5 provides conclusions.

1. Semantics of negative pronouns

A large number of studies have been devoted to the semantics of negative pronouns (see the links above). As E. V. Paducheva notes, in general, three variants of their analysis can be distinguished ^[6].

The first option: "a negative pronoun is a non–referential pronoun in a negative context: '(it is not true that) someone'" ^[6]. In other words, negative pronouns are synonymous with pronouns on something or something with an additional condition: they must appear in the context of negation. (3b) shows a simple formalization of this analysis using first-order predicate logic.

(3a) nobody ? anyone/someone

(3b) nobody, someone = ?x: P(x)

Those expressions whose semantics are formalized using the existence quantifier ?, we will call existential quantifiers in the future. For example, in the framework of the analysis under consideration, nobody and someone are existential quantifiers.

Let's consider a concrete example in which the predicate P is the predicate came:

(4a) No one came = Someone came

(4b) ?x: came(x) ‘there is such an x that x came.’

However, the sentence *nobody came, unlike someone came, is ungrammatical, because the pronoun nobody should be in a negative context. Adding a negation to this sentence gives the actual interpretation:

(5a) No one came

(5b) ?x: came(x) ‘There is no such x that x came.’

According to this version of the analysis, negative pronouns like nobody replace indefinite pronouns like someone, someone in negative sentences. The majority of researchers in formal semantics are in favor of this variant of analysis ^{[1],[10],[13],[14]}. In addition, this analysis is usually accepted by default in works devoted to negative pronouns, but not aiming at a direct analysis of their semantics. ^[3],[4],[7]. In this paper, we will present arguments against such an interpretation of the semantics of negative pronouns.

The second version of the analysis: "a negative pronoun can correspond to a universal status – when negation enters the scope of the quantifier of generality" (here we again use the formulation of E. V. Paducheva ^[6]). As E. V. Paducheva notes, in this case the logical equivalence is valid: ?x: P(x) ? ?x: P(x). For example, if there is no x such that x came (?x: P(x)), then it is true for any x that x did not come (?x: P(x)).

Thus, for example, the pronoun nobody is considered synonymous with the expression every person. The corresponding formalization using first-order predicate logic is indicated in (6b).

(6a) no one ? every person

(6b) nobody, every person = ?x: P(x)

Those expressions whose semantics are formalized using the generality quantifier ?, we will call universal quantifiers in the future. For example, within the framework of the analysis under consideration, no one and every person are universal quantifiers.

Consider an example similar to (4), in which the predicate P is the predicate came:

(7a) No one came = Every person came

(7b) ?x: came(x) ‘For every x it is true that x came.’

It is easy to see that the interpretations of (7a) and (4a) are radically different. However, again, a negative pronoun requires that a verb negation be present in the sentence. When it is added, (7a) becomes similar to (4a). Note that in this analysis the negation is attached not to the whole formula, but to the predicate P, in this case came, which gives the actual interpretation.

(8a) No one came.

(8b) ?x: came(x) ‘For every x it is true that x did not come.’

This version of the analysis, advocated in ^[11],[12], is less standard, which may be explained by the fact that it loses the analogy between negative and indefinite pronouns present in the first version of the analysis. One way or another, the same interpretation corresponding to the valid one is achieved in both approaches considered, which raises the question of which of these approaches is correct.

It should be noted that, as we said above, there is a third option for analyzing negative pronouns. It can be assumed that negation is contained in the semantics of these pronouns. Thus, the second, verbal negation is of a consistent nature. Although this version of the analysis is defended in ^[5],[6], its adoption generates numerous complications. At the moment it is the least standard, and in the future we will not consider it, see the discussion in ^{[9, p. 109]}.

Next, we will call the analysis of negative pronouns as existential quantifiers ?-theory (3), and their analysis as universal quantifiers – ?-theory (6). In order to determine which of the presented approaches is correct, it is necessary to consider the linguistic data that one approach is able to explain and the other is not able to explain. Such data, as will be shown below, include sentences in which negative pronouns are combined with the modifier almost. Before proceeding to the discussion of the relevant proposals, it is necessary to briefly consider the semantics of this indicator.

2. Semantics of almost

The indicator is almost considered by researchers as an "approximator", i.e. an indicator expressing the category of approximation (approximation). Analysis of approximators and, in particular, linguistic units synonymous with almost (Eng. almost, Ital. quasi, German. fast, etc.) devoted a large number of works in cognitive ^{[15],[16],[17],[18]} and formal linguistics ^{[13],[19],[20],[21],[22]}.

Most researchers agree that almost (and similar indicators in other languages) have two components of meaning. Firstly, it is actually an approximative component. For example, the sentencePetya almost climbed the mountain reports that Petya is close to climbing the mountain. In other words, in the actual world, a situation was reached that was close to the fact that Petya climbed a mountain. The works are devoted to a detailed semantic analysis of the approximation category ^{[15],[16],[17],[18]}, an informal understanding of approximation is sufficient for further discussion.

The second component of semantics is almost negative. For example, the sentencePetya almost climbed the mountain at the same time reports that the top of the mountain was not reached by Petya. There are two approaches to analyzing this component. The semantic approach considers it as part of its own meaning almost ^{[13],[15],[20],[22]}. On the other hand, according to the pragmatic approach, this component is deduced by the listener as an implicature ^[19],[21]. In particular, if the speaker reports that Petya approached the top of the mountain, based on general communicative principles (Maxim Gries ^[23]), the listener concludes that Petya did not reach the top of the mountain (otherwise the speaker would have reported it). We will consider the negative component as part of its own semantics almost, and not an implicature, which does not affect the analysis of combinations of the type almost nobody.

The following question is of key importance for the following arguments: what is the scope of the approximator almost? The first answer is: it almost interacts with the value of the expression to which it is attached. For example, in the sentencePetya solved almost ten problems the approximator modifies the interpretation of the component of ten or ten problems. In other words, almost roughly synonymous with about in the relevant meaning.

The second answer: almost interacts with the meaning of the entire sentence. In this case , in the sentencePetya solved almost ten problems the approximator modifies the interpretation of the rest of the sentence minus almost, i.e.Petya solved ten problems. With such an analysis, it almost reports that a situation close to the one described by the sentence has been achievedPetya solved ten problems. Two variants of the scope analysis are almost presented in (9).

(9) Petya solved almost ten problems.

(9a) ? ‘Petya solved some number of tasks close to 10’ (almost refers to the component of ten tasks)

(9b) = There was a situation close to the situation ‘Peter solved ten problems’ (almost applies to the entire sentencePetya solved ten problems)

Note that both variants of the analysis give an intuitively correct interpretation of the sentence. For example, if Petya solved about ten problems, then it is true that there was a situation close to ‘Petya solved (exactly) ten problems’, and vice versa.

The variant of the analysis illustrated in (9b) is accepted in the works ^{[10],[13],[22]}. Let 's formulate it more strictly using the concept of alternatives ^{[14],[21],[22]}.

(10) semantics [ almost q ]

(10a) Negative component, Neg(ation): it is not true that q.

(10b) Approximative component Prox(imity): there is an alternative q’ that is close to q and true.

As mentioned above, it is assumed that it almost approximates the whole situation, i.e. it has a scope over the entire proposition q. Consider the semantics of (10) by example. The components Neg and Prox are presented in (11a) and (11b), and the receiving interpretation in (11b). This interpretation corresponds to the expected one.

(11) Petya solved almost 10 problems = almost [ Petya solved 10 problems ]

(11a) Neg: It is not true that Petya solved 10 problems.

(11b) Prox: There was a situation q’, close to the fact that Petya solved 10 problems. For example, Petya solved 8 problems.

(11b) Neg & Prox: Petya didn't solve 10 problems, but he solved 8 problems.

In ^{[10, p. 239-240],[13, p. 232-244]} it is shown that in order to substantiate the ?-theory of negative pronouns (3), it is necessary to use the above analysis of the approximator almost. The following section is devoted to the discussion of this fact.

3. Combinations of the almost nobody type – an analysis option

In the introduction, we mentioned that it is almost not combined with all pronouns and quantifier words. For example, the sentence *almost someone came in Russian is ungrammatical. The approximator analysis presented in (10) and (11) should almost predict the ungrammaticity of such combinations. Let's take another look at the semantics of the pronoun someone. It is repeated in (12) with a slight change: ?1x instead of ?x. In other words, we will assume that someone means not just ‘there is x’, but ‘there is (at least) one x'. Thus, as alternatives to the pronoun someone, we will consider the expressions two people, three people, etc., which will allow us to more clearly formulate the content of the approximative component of semantics in almost the examples below.

(12) someone = ?1x: P(x)

Consider the analysis of an ungrammatical sentence * almost someone came. This sentence states that the situation q ‘someone came’ (i.e., ‘at least one person came’) did not take place (13b), but at the same time some close situation took place, for example, 3 people came (13b). Obviously, this leads to a contradiction, as shown in (13g).

(13) *Almost [_q someone came].

(13a) Statement q: ?1x: came(x) ‘Someone came.’

(13b) Neg: ?1x: came(x) ‘It is not true that someone came.’ (= no one came)

(13b) Prox: ?3x: came(x) ‘3 people came’ (let 3 be a close alternative to 1)

(13g) Neg & Prox: No one came (13b) & 3 people came (13b) (contradiction)

Thus, the analysis described in section 3 gives the desired result: it predicts that the sentence * almost someone came self-contradictory, and therefore ungrammatical. However, the question arises – why, in this case, is the offer acceptable almost no one came? In this sentence, contradiction does not arise due to negation. In particular, it is claimed that the situation "no one came" did not take place, respectively, at least one person came (14b). At the same time, it is claimed that there was some kind of situation close to ‘no one came’, for example, ‘only 2 people came’ (14b). Thus, the actual interpretation (14g) is obtained.

(14) Almost [_q no one came ].

(14a) q: ?1x: came(x) ‘It is not true that someone came.’ (= no one came)

(14b) Neg: ?1x: came(x) = ?1x: came(x) ‘Someone came.’

(14b) Prox: ?3x: came(x) ‘It is not true that 3 people came.’

(14g) Neg & Prox: someone came (14b) & it is not true that 3 people came (14b) = only 2 people came

Since in the framework of the ?-theory, the pronouns nobody and someone are analyzed as synonyms (3), examples (13) and (14) differ only in the presence of negation in the sentence. Thus, for the presented analysis, it is fundamentally important that the approximator almost refers not to a single word (someone and no one), but to the entire proposition, including negation. However, in the next section we will provide data that calls into question such an analysis almost.

4. Combinations of almost pronouns and numerals – new data

If we consider no one as a synonym for the indefinite pronouns someone, someone or anyone, then these pronouns should exhibit similar properties in the same contexts. For example, the sentences in (15) taken from ^{[5, p. 224]} appear to be synonymous (both of these sentences are also found in large numbers on the Internet).

(15a) There is no evidence.

(15b) There is no evidence.

However, when added, almost only sentence (15a) remains grammatical.

(16a) There is almost no evidence.

(16b) *There is almost no evidence.

From the difference in (16), it can be concluded that the pronouns nobody and anyone (as well as someone and someone) have different semantics, which is an argument against the ?-theory considering them as synonyms. Note that everyone (17) is almost freely combined with the quantifier word, which is confirmed by the analysis considering nobody and everyone as synonyms (6).

(17) Almost everyone knows that …

This problem for ?-theory was noticed earlier in ^[13],[22] when discussing English language data. In this paper, we would like to give another argument against the ?-theory. As mentioned above, it is fundamentally important for this theory that the approximator almost does not refer directly to the negative pronoun (for example, no one), but to the entire sentence (for example, no one came (14)). However, such an analysis of the scope almost predicts an incorrect interpretation of sentences with numerals.

Consider the offerPetya has not solved almost 10 problems. Suppose that almost applies to the entire sentence (18). In this case, the negative component states that the situation is q ‘Petya did not solve 10 problems’ did not take place, i.e. due to the double negation, it is true that Petya solved 10 problems (18b). In addition, there was some situation close to q, for example, ‘Peter did not solve 12 problems' (18b). Thus, the sentence should be interpreted as ‘Peter solved 10 or 11 problems’ (18g), but there is no such interpretation for sentence (18). In particular, the discourse presented in (19), which corresponds to this interpretation, was considered anomalous by almost all the media we interviewed.

(18) Petya has not solved almost 10 problems = almost [_q Petya has not solved 10 problems ]

(18a) q: it is not true that Petya solved 10 problems

(18b) Neg: (it is not true that Petya solved 10 problems) = Petya solved 10 problems.

(18b) Prox: it is not true that Petya solved 12 problems (let 12 be a close alternative to 10)

(18g) Neg & Prox: Petya solved 10 problems (18b) & it is not true that Petya solved 12 problems (18b) = Petya solved 10 or 11 problems.

(19) *Petya has not solved almost 10 problems. To be precise, he solved 11 problems.

5. Discussion and conclusions

In section 3, we considered two options for analyzing the approximator almost. In the first case, it modifies the value of the component to which it adjoins. In this case, for example, the sentencePetya solved almost 10 problems means ‘Petya solved some number of problems close to 10'. On the other hand, it can be assumed that almost applies to the entire sentence. ThenPetya solved almost 10 problems means ‘There was a situation close to the situation in which Petya solved 10 problems'. As shown in section 3, only the second version of the analysis is compatible with the ?-theory of negative pronouns. However, this approach cannot explain why it is almost combined with nobody, but not combined with its supposed synonyms anyone, someone, anyone. In addition, we have shown that this approach predicts incorrect interpretations in sentences with numerals and negation (Section 4).

Thus, we come to the conclusion that the approximator almost refers to a separate component, and not to the entire proposition. Consequently, contrary to the opinion of some authors ^[10],[13] the compatibility with is almost relevant for the semantic analysis of negative pronouns. Since it is almost combined with both the quantifier word each (17) and negative pronouns, we come to the conclusion that the ?-theory considering negative pronouns as universal quantifiers (6) is correct.

Thanks

The study was supported by a grant from the Russian National Science Foundation 22-18-00285.