Skip to content
BY 4.0 license Open Access Published by De Gruyter Mouton September 22, 2020

DP-internal Inversion and Negative Polarity: Latin aliquis and its Romance Descendants

Chiara Gianollo EMAIL logo
From the journal Probus

Abstract

I analyze the Romance descendants of Latin aliquis ‘some or other’, which are characterized by a complex pattern of variation in the contemporary Romance languages. I account for this variation in terms of diverging diachronic paths, tracing their determinants back to a process taking place between Classical and late Latin. Classical Latin only used aliquis as an epistemic indefinite, expressing ignorance about the identity of the referent. In late Latin a distributional extension is observed, and aliquis starts to be consistently found as an NPI in negative contexts. This multiplicity of uses is transmitted to medieval Romance and represents the prerequisite for contemporary variation. In their further history, some languages continue only one of the two uses. Other languages maintain both, but the meaning contrast comes to be related to a word-order difference. I analyze this difference as a syntactic DP-internal inversion operation, motivated by focus and connected to polarity sensitivity. Significantly, the diachronic path of the Romance descendants of aliquis contributes to our understanding of general mechanisms of semantic change, since it instantiates a cline of development that can be related to varying (hence, diachronically changing) constraints on quantificational domains.

1 Introduction

European Portuguese shows a major contrast in the interpretation of the indefinite determiner algum, which appears to be connected to its position within the nominal phrase. The relevant paradigm, first investigated by Martins (2015a, 2015b), is presented in (1): the indefinite determiner algum is interpreted as ‘some’ in (1a), where it precedes the head noun, whereas it is interpreted as ‘no’ in (1b), where it follows the noun:

(1)

European Portuguese (Martins 2015a: 400)

a. Algum lugar se parece com este
some:sg place:sg refl seem:3sg like this:sg
‘There is some place that looks like this one’
b. Lugar algum se parece com este
place:sg no:sg refl seem:3sg like this:sg
‘There is no place that looks like this one’

It seems, thus, that Portuguese algum can be either “positive” or “negative”, and that this interpretive distinction correlates with a word-order difference within the nominal phrase.[1]

A similar, although not completely equivalent, interpretive contrast is observed in Spanish: the cognate indefinite determiner algún precedes the noun and is interpreted as ‘some’ in (2a), which is a positive-polarity assertion; in negative-polarity contexts, instead, it is interpreted as ‘any’ and it necessarily follows the noun (2b–c).

(2)

Spanish[2]

a. Me he encontrado con algún estudiante
refl have:1sg met with some:sg student:sg
‘I met some student’
b. No me he encontrado con estudiante alguno
not refl have:1sg met with student:sg any:sg
‘I did not meet any student’
c. * No me he encontrado con algún estudiante
not refl have:1sg met with any:sg student:sg

In Italian and French no such contrast obtains: the cognate determiners alcuno and aucun in the singular only have the meaning ‘any’ and are found exclusively in negative-polarity environments, cf. Italian in (3a–b).[3] Inversion with respect to the noun is optional and has an emphatic flavor.

(3)

Italian

a. Non ha pagato le tasse alcuno studente
not have:3sg paid the:pl fees:pl any:sg student:sg
‘No student paid the fees’
b. * Alcuno studente ha pagato le tasse
any:sg student:sg have:3sg paid the:pl fees:pl
c. Non ha pagato le tasse studente alcuno
not have:3sg Paid the:pl fees:pl student:sg any:sg
‘No student (at all) paid the fees’

We observe thus a complex pattern of variation across Romance in the behavior of elements of the functional lexicon that share a common Latin origin, namely the Latin indefinite determiner aliquis ‘some or other’ (which in its Romance descendants is invariably compounded with unus ‘one’).

This pattern raises a number of questions that are significant for comparative Romance linguistics and historical linguistics alike: which factors are responsible for the pertinacity and, at the same time, the evolution of the Romance descendants of Latin aliquis? How is it possible that an originally positive determiner (‘some or other’) develops a negative meaning (‘any’, ‘no’)? How can the two meanings coexist in the same language? How can the role of word order in the interpretation be accounted for by current models of the syntax-semantics interface?

In this contribution I set out to propose answers to these questions. In §2 I provide a more detailed picture for the distribution of the outcomes of aliquis in the standard Romance languages. In the rest of the paper I attempt an explanation for the observed synchronic variation in terms of the diachronic path followed by the Latin ancestor and by its continuations in the various languages. Section 3 presents the main features of Classical Latin aliquis and follows its development in late Latin. Section 4 turns to the medieval Romance languages and singles out aspects of continuity with respect to the Latin situation, as well as elements of differentiation across Romance. In §5 I draw some general conclusions on the diachronic cline, and I discuss the role of DP-internal inversion and its grammaticalization in Ibero-Romance. Section 6 concludes the paper summarizing the main conclusions.[4]

By means of this study I hope to show that the history of Latin aliquis and its descendants provides evidence for systematic processes of semantic change involving the functional lexicon. The relevant dimension of variation is represented by the constraints that quantificational determiners impose on their domain of quantification. Semantic change targets these constraints in situations where additional inferences are licensed by interacting operators in the clause. In the course of time, these additional contextual inferences may become reanalyzed as an obligatory meaning contribution of the determiner itself, that is, as coded in its lexical entry. This, in turn, may affect the syntactic component: in the case at hand, the result is the conventionalization of an originally optional movement operation (DP-internal inversion) and the development of formal features inducing a Negative Concord relation with negation.

2 The Synchronic Distribution

2.1 Forms and Overview of Distributional Tendencies

In this section I analyze the distribution of the indefinite determiners that continue Latin aliquis in the following standard Romance languages: Italian (alcuno), French (aucun), Catalan (algun), Spanish (algún), European Portuguese (algum).[5] All these forms derive from the combination of Latin aliquis ‘some or other’ with the cardinal numeral unus ‘one’: aliquis unus > *alicunus (REW, 1935 §339). The combination as such is unattested in Latin texts and has to be reconstructed for an undocumented pan-Romance stage. Some languages also have pronominal items that continue the simple form (masc. acc. aliquem, neuter nom.–acc. aliquid), such as Sp. alguien ‘someone’, Pt. alguem ‘someone’, Sp. and Pt. algo ‘something’, ‘a bit’ (REW, 1935 §345).

In my overview I concentrate on the quantificational determiner, which exhibits the relevant variation in word order with respect to the head noun. The main aspects of import for Romance historical linguistics, and more in general for theoretical approaches to language change, are the following:

  1. In all the investigated languages except contemporary Catalan the singular can be used in negative environments, unlike Classical Latin aliquis; the licensing relation is, however, not the same in all languages;

  2. In some languages the negative use coexists with a positive one; however, the two uses are distinguished by word order, since in the negative use the determiner must follow the noun.

  3. In all the investigated languages, the singular and the plural forms of the determiner have more or less diverging distributional patterns, to such an extent that it is not always clear that they can be treated as belonging to a single lexical item.

In order to keep the discussion to a manageable length, in the following sections I will deal with aspects (i) and (ii), which concern the singular, but will have to put the plural aside. All the investigated languages also continue the plural form. In contemporary French it is residual (confined to archaic d’aucuns ‘some (pl.)’ and to pluralia tantum). In the other languages under scrutiny it has a productive use as a plain indefinite that alternates with plural forms of the indefinite article (Crisma 2012; Martí 2008).[6] The plural form tends to receive wide scope in the interaction with other operators. In Italian, in particular, it resists taking narrow scope with respect to negation (4a), despite the fact that the singular is systematically interpreted in the scope of negation (4b).

(4)

Italian

a. Francesco non ha considerato alcuni commenti
Francesco not have:3sg considered some:pl comment:pl

‘Francesco did not consider some comments’ = there were some comments that Francesco did not consider

b. Francesco non ha considerato alcun commento
Francesco not have:3sg considered any:sg comment:sg

‘Francesco did not consider any comment’

In Portuguese the plural is impossible with inverted algum (Martins 2015b: 105). In Spanish the singular form is strongly preferred in negative contexts. If the plural is used, it has to follow the head noun, similarly to the singular.

An important difference with respect to the singular form in positive contexts is that the plural does not trigger the same epistemic inferences with respect to the identity of the referents that are observed with the singular. For attempts to reconcile the singular and plural uses under a single lexical entry, see Alonso-Ovalle and Menéndez-Benito (2011) and Etxeberria and Giannakidou (2017).

In the singular, three main distributional patterns can be distinguished (5):[7]

(5)

Distributional patterns across Romance for the singular form

a. Italian, French: only in negative contexts
b. Catalan: only in positive contexts
c. Spanish, European Portuguese: both in positive and negative contexts

2.2 The Epistemic Use

Let us start with the more complex pattern of Spanish and Portuguese, which is also the best studied. A precise characterization of the positive-polarity contexts of occurrence comes from work on Spanish by Alonso-Ovalle and Menéndez-Benito (2003, 2010, who define algún as an epistemic indefinite. Epistemic indefinites are existential determiners that indicate that the epistemic agent to which the indefinite is anchored (the speaker or the attitude holder) is ignorant or indifferent with respect to the identity of the referent.[8]

In (6) the referent of the DP containing algún cannot be further specified, as shown by the possibility of a continuation that makes the ignorance component explicit (6a), by the impossibility of further specifying the intended referent (6b), and by the impossibility of an anaphoric relation (6c):

(6)

Spanish

a. Algún estudiante pagó las tasas de matrícula
some:sg student:sg pay:3sg the:pl fee:pl of tuition:sg
pero no quién es
but not know:1sg who be:2sg
‘Some student paid the tuition fees, but I don’t know who s/he is’
b. María está buscando a algún estudiante.
María stay:3sg looking at some:sg student:sg
# Su nombre es Juan
his:sg name:sg be:3sg Juan
‘Mary is looking for some student. His name is John’
c. Algún estudiante pagó las tasas de matrícula.
some:sg student:sg pay:3sg the:pl fee:pl of tuition:sg
# Hablé con él después
speak:1sg with him afterwards
‘Some student paid the tuition fees. I spoke to him afterwards’

The paradigm can be replicated for Portuguese (7):

(7)

Portuguese

a. Algum estudante pagou as propinas.
some:sg student:sg pay:3sg the:pl fee:pl
Mas não sei quem (foi)
but not know:1sg who be:3sg
‘Some student paid the tuition fees, but I don’t know who s/he is’
b. A Mary está à procura de algum estudante.
the:sg Mary stay:3sg to search:sg of some:sg student:sg
# O nome dele é John
the:sg name:sg of.he be:3sg John
‘Mary is looking for some student. His name is John’
c. Algum estudante pagou as propinas.
some:sg student:sg pay:3sg the:pl fee:pl
# Falei com ele depois
speak:1sg with him afterwards
‘Some student paid the tuition fees. I spoke to him afterwards’

The ignorance component results in non-specific interpretations and, in particular, in obligatory narrow-scope readings with respect to intensional operators (de dicto readings, cf. 6b and 7b).[9]

In order to account for the ignorance component, I follow Alonso-Ovalle and Menéndez-Benito (2010), who treat it as a conversational implicature, triggered by a constraint on the quantificational domain of epistemic indefinites. According to this constraint (anti-singleton presupposition in Alonso-Ovalle and Menéndez-Benito 2010: 15–19), the domain of quantification of algún cannot be narrowed down to a singleton, contrary to what happens with specific indefinites (Schwarzschild 2002). This results in minimal domain widening: the quantificational domain must include more than one individual. At the same time, there is no instruction to maximally widen the domain as to include even marginal individuals: this represents an important difference with respect to free-choice indefinites (‘whoever’) and negative polarity items (‘any’).

Epistemic indefinites are, thus, ‘intermediate’ creatures between specific (singleton) indefinites and maximal domain wideners. Their felicity depends on the compatibility of the inferences they trigger with the broader context. The ignorance implicature emerges once the epistemic indefinite enters a competition with determiners imposing smaller domains (definites, specific indefinites). In order to be felicitous, the pragmatic contribution has to serve a communicative need; Alonso-Ovalle and Menéndez-Benito (2010), following Kratzer and Shimoyama (2002) identify two main pragmatic functions that grant a felicitous use of epistemic indefinites: the need to avoid making a false claim and the need to avoid exhaustivity inferences. For this reason, epistemic indefinites are felicitous in (explicitly or implicitly) modalized contexts, where the weaker statement provided by epistemic indefinites is pragmatically motivated; episodic contexts like (6a, c, 7a, c) are considered implicitly modalized due to the presence of a covert assertoric operator (Alonso-Ovalle and Menéndez-Benito 2010: 7–8).

Although specific studies are lacking, Catalan algun can also be classified as an epistemic indefinite based on the tests used for Spanish and Portuguese: the ignorance component can be made explicit (8a), no further specification of the intended referent is possible (8b), and no anaphoric relation can be established (8c):

(8)

Catalan

a. Algun estudiant va pagar la matrícula,
some:sg student:sg go:3sg pay the:sg fee:sg
però no qui és
but not know:1sg who be:3sg
‘Some student paid the tuition fees, but I don’t know who s/he is’
b. La Maria busca algun estudiant.
the:sg Mary look.for:3sg some:sg student:sg
# Es diu Joan
refl say:3sg Joan
‘Mary is looking for some student. His name is John’
c. Algú es deu haver oblitad de pagar
someone:sg refl must:3sg have forgotten of pay
la matrícula. # L’ he vist al despatx
the:sg fee:sg him have:1sg seen at.the:sg office:sg
‘Someone must have forgotten to pay the tuition fees. I saw him in the office’

In Catalan, only the epistemic use is possible in the singular. In Spanish and Portuguese, on the other hand, the determiner also has special uses in negative contexts, which will be the object of §2.3.

2.3 The Negative Contexts

Turning to the negative contexts, we see that Spanish algún and Portuguese algum can be found as existential indefinites taking narrow scope with respect to sentential negation (to which I restrict my attention), cf. (1b) and (2b). In both languages, the determiner must follow the noun when interpreted under sentential negation (cf. 9 for Portuguese, and 2c for Spanish).[10]

(9)

Portuguese

a. Eu não encontrei estudante algum
I not meet:1sg student:sg any:sg
‘I did not meet any student’
b. Eu não encontrei algum estudante
I not meet:1sg any:sg student:sg

Despite the overall similarity, an interesting contrast between the two languages can be observed: in Portuguese, algum has the distribution of a Negative Concord Item (NCI or n-word), whereas Spanish algún behaves as a Negative Polarity Item (NPI). Following Zeijlstra (2004), Penka (2011), I assume that the two classes of items are subject to different licensing requirements (cf. §5.2 for further discussion). NPIs are subject to a semantic-pragmatic felicity condition that requires them to occur in the scope of overt operators creating downward-entailing contexts.[11] NCIs undergo a syntactic mechanism of Agree with a c-commanding negative operator, induced by the formal uninterpretable features they carry. In the case of NCIs the licensing operator can be covert: this is the case when NCIs precede the verb (hence, also the negative marker) or are found in isolation as negative short answers: in those cases, they appear to be “self-licensing” (Ladusaw 1992), since they are the only overt elements expressing negation.

Portuguese algum can be found in pre-verbal DPs; when pre-verbal, it negates the sentence by itself (10a), and co-occurrence with the negative marker is excluded in a single-negation reading. Moreover, it can be used as a negative short answer (10b), a possibility that is precluded with NPIs, which need an overt, c-commanding licenser.

(10)

Portuguese

a. Estudante algum pagou as propinas
student:sg any:sg pay:3sg the:pl fee:pl
‘No student paid the tuition fees’
b. Speaker A: Vais vender a tua casa?
go:2sg sell the:sg your:sg house:sg
‘Are you going to sell your house?’
Speaker B: Em circunstância alguma
in circumstance:sg any:sg
‘Under no circumstance’ (Martins 2015a: 407)

Spanish algúm, on the other hand, is only found in post-verbal position in negative sentences, where it is c-commanded by the overt negative marker: it is impossible in pre-verbal subject position and as a negative short answer (11a–b).

(11)

Spanish

a. * Estudiante alguno no ha pagado las tasas
student:sg any:sg no have:3sg paid the:pl fee:pl
de matrícula
of tuition:sg
‘No student paid the tuition fees’
b. Speaker A: ¿Tienes objeciones a eso?
have:2sg objection:pl to this:sg
‘Do you have objections to this?’
Speaker B: * Objeción alguna
objection:sg any:sg
‘No objection’

To summarize, in Spanish and Portuguese the singular form occurs both as an epistemic indefinite and as an existential indefinite with narrow scope with respect to negation; crucially, the two uses are syntactically disambiguated by word order: in the scope of negation, inversion between the head noun and the indefinite is obligatory.

In French and Italian, the continuants of Latin aliquis in the singular are only possible in negative contexts (and residually in some non-negative NPI-licensing contexts: cf. Crisma 2012: 514–517; Milner 1979: 81–82). A difference in distribution is also observed between these two languages: French aucun is an NCI, whereas Italian alcuno is an NPI. The distinguishing contexts are parallel to those seen for Portuguese and Spanish. French aucun can occur in pre-verbal position and as a negative short answer (12a–b); in the standard variety it always co-occurs with ne and it can be found in a Negative Concord configuration with another n-word (negative spread, 12c).

(12)

French

a. Aucun étudiant n’ a payé les frais d’ inscription
no:sg student:sg not have:3sg paid the:pl fee:sg of tuition:sg
‘No student paid the tuition fees’
b. Speaker A: Avez-vous une objection ?
have:2pl-you:pl a:sg objection:sg
‘Do you have an objection?’
Speaker B: Aucune objection
no:sg objection:sg
‘No objection’
c. Personne n’ a vu aucun film de Bresson
nobody:sg not have:3sg seen no:sg film:sg of Bresson
‘Nobody has seen any film by Bresson’ (Corblin 1996: ex. 88)[12]

Italian alcuno cannot precede the verb and cannot be used as negative short answer (13a–b; cf. also 3b), meaning that it is never “self-licensing”. It has to be licensed by a hierarchically superior negative marker (13c) or n-word (13d).

(13)

Italian

a. * Alcuna obiezione è stata espressa
any:sg objection:sg be:3sg been expressed
durante la discussione
during the:sg discussion:sg
‘No objection was raised during the discussion’
b. Speaker A: Hai obiezioni?
have:2sg objection:pl
‘Do you have objections?’
Speaker B: * Alcuna obiezione
any:sg objection:sg
‘No objection’
c. Non è stata espressa alcuna obiezione
not be:3sg Been expressed any:sg objection:sg
durante la discussione
during the:sg discussion:sg
‘No objection was raised during the discussion’
d. Nessun partecipante ha espresso alcuna obiezione
no:sg participant:sg have:3sg expressed any:sg objection:sg
‘No participant expressed any objection’

Inversion with respect to the head noun is possible, but has no effect on the distribution seen above (14–15). Speakers judge the inverted orders as more emphatic (cf. 3c) and typical of a formal, conservative style; especially in French inversion is a very marginal option, basically occurring only in fixed expressions after sans ‘without’.

(14)

French

soit dit sans offense aucune
be:3sg said without offense:sg no:sg

‘let it be said without any offense’

(15)

Italian

Certe persone non mostrano rispetto alcuno
some:pl people:pl not show:3pl respect:sg any:sg
per le istituzioni
for the:pl institutions:pl

‘Some people do not show any respect for the institutions’

To summarize, in French and Italian the singular form is restricted to negative contexts and is interpreted as an existential taking narrow scope with respect to negation; in Italian it has the distribution of a Negative Polarity Item (NPI), in (standard) French of a Negative Concord Item (NCI or n-word).

2.4 Summary

Table 1 provides an overview of the data discussed in this section.

Table 1:

Summary for uses in the singular.

Only in scope of negation Only epistemic indefinite In scope of negation & epistemic indefinite
Italian alcuno Catalan algun Spanish algún
French aucun Portuguese algum

Concerning inversion in the order of the indefinite determiner and the head noun, the main observation is that in some Romance languages (Spanish, Portuguese) it is apparently related to a positive / negative flip and is obligatory in the relevant contexts, whereas in other Romance languages (French, Italian) it is an optional, stylistically marked operation with limited import on interpretation.

Catalan is the only language lacking uses in the scope of sentential negation. However, as we will see in §4, earlier stages of the language witness an additional NPI use, which was subsequently lost: we can therefore conclude that in all the standard Romance languages the continuants of aliquis are found in negative contexts at some point in history.

3 Latin aliquis

This section deals with the Latin ancestor, aliquis, with the aim of understanding the point of departure for the variation observed in Romance. At all stages of Latin, aliquis ‘some or other’ (< stem of alius ‘other’ + interrogative / indefinite stem *k w i-) can be used either as a pronoun or as a quantificational determiner, both in the singular and in the plural. It is more often prenominal, but inversion with respect to the head noun is possible (Marouzeau 1922: 165–166).

The first aspect I will address is whether the positive use of aliquis is comparable to the positive use in the contemporary Romance languages, which I defined as epistemic. I argue that early and Classical Latin aliquis can indeed be characterized as an epistemic indefinite, and that this use also characterizes late Latin.[13] However, the investigation of the late Latin data will lead us to deal with a further aspect, which has to do with the expansion of aliquis into negative contexts in late Latin. I will discuss some bridging contexts that allow us to see how the epistemic determiner comes to be interpreted as a maximal domain widener in the scope of downward-entailing operators. This way, a substantial continuity between late Latin and Romance will emerge, as we will further see in §4, dedicated to medieval Romance.

Early and Classical Latin singular aliquis is regularly associated with an ignorance inference: the speaker cannot, or does not want to, further identify a referent. According to Bortolussi (2015: 24), aliquis imposes a variation requirement on its quantificational domain: it entails that there exists a set of individuals who are possible denotational targets, with no commitment to unique identifiability. We see that this description matches the behavior of epistemic indefinites.

The sentence in (16) exemplifies the non-identifiability of the referent:

(16)

early Latin (Plaut. Capt. 382–383)

nam pater exspectat aut me aut aliquem nuntium
indeed father:nom wait:3sg or me or some:acc messenger:acc
qui hinc ad se veniat
who:nom from.here to refl come:3sg

‘your father is expecting either me or some other messenger who is coming to him from here’

In (16), aliquis is the object of an intensional verb, a context where lack of identifiability is felicitous. In fact, occurrences in episodic, realis contexts are very rare. The most typical contexts for aliquis are explicitly or implicitly modalized utterances, as (17), where the verb in the subjunctive expresses epistemic possibility.

(17)

Classical Latin (Cic. Verr. 2, 2.78)

forsitan aliquis aliquando eiusmodi quippiam fecerit
maybe someone:nom sometimes of.this.sort something:acc do:3sg

‘there may now and then have been someone who did something of the same description’

The conclusion, reached in §2, that all the standard Romance languages (at some stage) use the descendants of aliquis in negative contexts is in principle surprising in view of the fact that the Classical Latin ancestor was rarely found in the context of sentential negation, and in the few cases where it is found there it escaped the scope of the negation operator. The example in (18) shows this particularly clearly:

(18)

Classical Latin (Sen. ep. 78.19)

inter haec tamen aliquis non gemuit
among this:acc though someone:nom not lament:3sg

‘Among all this (these tortures) there has been someone, though, who did not lament’

Here the referent is left unspecified (and in fact the context suggests that there may have been multiple people resisting torture without lamenting); despite the non-specific reading, there is an existential entailment, showing that aliquis is interpreted outside the scope of negation.

The passage from Lucretius in (19), where aliquis is found in the object DP of a negative sentence (a very rare configuration in Classical Latin), is more complex.

(19)

Classical Latin (Lucr. 6, 674–675)

scilicet et fluvius qui visus maximus
obviously also river:nom which:nom seen:ptcp.nom large:sup.nom
ei qui non ante aliquem maiorem
he:dat who:nom not before some:acc big:comp.acc
vidit
see:3sg

‘obviously such [i.e., considered huge] will also be a river, which seems the largest to a person who has not seen a larger river before’

The broader context indicates that aliquis in this passage has the value of ‘a, some’, not of ‘any’: the domain-widening component is absent, and the existence of a representative for the DP denotation (i.e., a larger river) needs to be presupposed in the argumentation. Lucretius is defending the thesis that volcanic eruptions are natural phenomena, not caused by divine intervention. To the objection that they are too huge to be caused by nature alone, Lucretius answers that this is something that humans may think just because they only know their world and they imagine as huge all the biggest things that they have seen; they do not understand that their world is nothing compared to the universe (vv. 676–679). He uses the simile in (19): a man considers a river he sees as the largest, because he never saw a larger one. Clearly, for the reasoning to go through, the presupposition that indeed there exist larger rivers (that the man has not seen) must hold. Although the reading is not specific (there is no reference to a determined specimen), the existence of the entity is not denied.

In Classical Latin aliquis is used rarely also in downward-entailing contexts other than negation; when it is found there, aliquis is not a maximal domain widener (cf. Devine and Stephens 2013: 378; Gianollo 2018: 77; Traina and Bertotti 1985: 189). In this, it contrasts with indefinites like quisquam and ullus, which are equivalent to ‘any’.

The situation changes in late Latin. Here I focus on the language of Augustine, which is particularly innovative in many respects concerning quantificational elements; however, a similar situation is found also in other contemporary and later texts that use a non-conservative, colloquial style, mirroring the ongoing changes.

Augustine still uses aliquis as an epistemic indefinite: in (20) he wants to leave the denotation of the book undetermined, because it still has to be written.

(20)

late Latin (Aug. ep. 246.3)

his itaque paucis si aviditas tua contenta
this:abl then few:abl if eagerness:nom your:nom satisfied:nom
non est et librum aliquem de hac
not be:3sg and book:acc some:acc about this:abl
re, quem diutius legas, desideras,
thing:abl which:acc longer read:2sg wish:2sg
patienter tibi sunt expectandae vacationes
patiently you:dat be:3PL expect:ger free.time:nom
nostrae
our:nom

‘So then, if these few remarks do not appease your thirst for information and you desire on this subject a book that will take a longer time to read, you must await with patience until I have a free interval’

However, Augustine also uses aliquis innovatively in the scope of sentential negation, and does so quite frequently: I present two such cases in (21) and (22), which are simple structures whose reading unambiguously requires narrow scope with respect to negation.

(21)

late Latin (Aug. in Ps. 93.7)

tetigit omnes, non timuit aliquem
touch:3sg everyone:acc not fear:3sg anyone:acc

‘He (Jesus) touched everyone, he did not fear anyone’

(22)

late Latin (Aug. in Ps. 99.13)

praeter uxorem suam non sciens aliquam
except wife:acc his:acc not knowing:nom anyone:acc
praeter maritum suum non sciens aliquem
except husband:acc her:acc not knowing:nom anyone:acc

‘(the chosen will be) he who does not know any other woman except his wife, she who does not know any other man except her husband’

Also in downward-entailing contexts other than negation aliquis is used in a way that suggests the possibility of maximal domain widening (excluding all the possibilities); (23) is a case in which aliquis occurs twice: the first instance is clearly an epistemic indefinite, but the second one, which is embedded in the standard of comparison (a downward-entailing context) is ambiguous between a minimally widened (‘someone else’) and a maximally widened (‘anyone’) interpretation.

(23)

late Latin (Aug. serm. 65A, 11)

non solum enim peccatum est diligere
not only indeed sin:nom be:3sg love:inf
aliquem plus quam Christum, sed peccatum
someone:acc more than Christ:acc but sin:nom
est non plus quam aliquem diligere
be:3sg not more than someone:acc love:inf
Christum
Christ:acc

‘not only it is a sin to love someone more than Christ, but it is a sin to love Christ not more than someone else / anyone’

Cases like (23) may well constitute bridging contexts in the semantic development, where both interpretations are equally possible and contextual inferences may lead to favor the innovative interpretation over the conservative one. These contexts suggest that the extension from epistemic into NPI environments is a diachronically natural development (a step from minimal to maximal domain widening, as we will see in §5). Non-negative NPI-licensing contexts were available to aliquis already in early and Classical Latin, and are good candidates as the locus for the semantic reanalysis. In turn, the rise in frequency of readings with maximal domain widening in these contexts may have triggered the further extension into negative contexts (with the relevant, narrow-scope reading seen in (21) and (22)). Additional conditions favoring the extension must be due to co-occurring systemic changes (for instance, the removal of blocking effects caused by other determiners like nullus ‘no’, changes in word order) in ways that should be investigated by further research (cf. Gianollo 2018: Ch. 5 for an attempt in this direction).

In conclusion, aliquis was an epistemic indefinite in early and Classical Latin and continues to be used as such in late Latin. In late Latin, however, an extension into negative contexts starts to appear. In §4 we will see that all the medieval Romance languages continuing aliquis had both the epistemic use and the uses in negative contexts. Thus, once we consider the developments taking place in late Latin, we can treat the whole range of uses displayed by the medieval Romance continuants of aliquis as a common inheritance.

4 The Medieval Romance Descendants

This section provides a compact description of the distribution of the descendants of aliquis in the earliest attestations of the Romance languages under investigation. The comparative overview will provide exemplification for the following points, which are particularly significant in a diachronic perspective:

  1. All languages had both the positive and the negative use;

  2. Non-negative NPI uses were more frequent in all medieval varieties than in the contemporary ones;

  3. Inversion was possible in negative contexts, but never obligatory, and it was an option (‘free inversion’ or DP-level scrambling in Martins 2015b: 118) available with all quantifying determiners.

The earliest stages of old French are characterized by a conservative retention of all uses also seen for late Latin. However, aucun is rare in negative contexts at the earliest documented stages. An example is provided in (24):

(24)

old French (Le Jeu de la feuillée 670–671, from Foulet 1970: §358)

Dame, or ne faites tel despit /
Milady now not make:2pl such contempt
k’ il n’ aient de vous
that they:pl not have:3pl from you:pl
aucun bien
any:sg good:sg

‘Milady, do not be upset to such a point that they will not receive from you any good’

Sentential negation is much more often accompanied by nul, which continues Latin nullus ‘no’, and has NCI as well as NPI uses. Ingham (2011), Ingham and Kallel (2014) follow in detail the process of replacement. They provide abundant documentation on early NPI uses of aucun in non-negative contexts (especially in the protasis of conditional clauses). No existing study focuses on the epistemic use, but in fact many of the contexts that are characterized as “positive” are compatible with an analysis of aucun as an epistemic indefinite. An example is (25):

(25)

old French (Béroul 111.3627)

Qant aucun passe devant lui, En plaignant
when someone:sg pass:3sg in.front.of him in crying
disoit: “Mar i fui! […]”
say:3sg poor I be:1sg

‘When someone would walk by him, he would say crying: “Poor me!”

In the diachronic development, aucun completely substitutes nul in negative contexts. However, it is superseded by a newly grammaticalized determiner, quelque, in epistemic and NPI contexts (Foulet 1919; Jayez and Tovena 2011). In the history of French aucun, we thus see a clear shift from positive to negative contexts, which, according to Ingham and Kallel’s (2014) data, is concluded by the beginning of the 18th century.

Old Italian presents a similar situation: epistemic uses are very frequent in the early attestations, together with NPI uses; the latter, besides occurring in the scope of sentential negation, are particularly widespread in non-negative NPI contexts (Stark 2002, 2006). The epistemic component of alcuno makes it particularly suitable for expressing background information, in contexts where identifiability of the denotation is not required and no discourse-prominent referent is introduced (Ledgeway 2009: 187–188 for old Neapolitan; Stark 2002: 326–327 for old Tuscan). The diachronic study in Stark (2002, 2006 shows an increasing preference for alcuno in negative contexts in old Tuscan from the 13th to the 15th centuries.

Below I provide examples for the epistemic use (26), for the NPI use in the conditional protasis (27), and for the NPI use under negation (28) from different early Italo-Romance varieties.

(26)

old Tuscan (Brunetto Latini, Rettorica 138.2)

è contra naturale usanza ch’ alcuna femina
be:3sg against natural:sg course:sg that some:sg woman:sg
uccida il suo marito e figliuoli e
kill:3sg the:sg her:sg husband:sg and child:pl and
distrugga un alto reame
destroy:3sg a:sg prestigious:sg realm:sg

‘it is against the natural course of events that some woman kills her husband and children and destroys a prestigious realm’

(27)

old Venetian (Statuti veneziani 152.70.35)

Et se alcun deli predicti ha
and if any:sg of.the:pl above-mentioned:pl have:3sg
alcuno officio al presente, perda quello
any:sg appointment:sg at.the:sg present:sg lose:3sg that:sg

‘And if any of the above-mentioned has any appointment at the moment, he shall lose it’

(28)

old Neapolitan (Libro de la destructione de Troya 314.15 P, from Ledgeway 2009: 188)

non facendo alcuna mentione dello cavallo predicto
not making any:sg mention:sg of.the:sg horse:sg above-mentioned:sg

‘not making any mention of the above-mentioned horse’

Examples (27)-(28) show that inversion was not necessary in negative and non-negative NPI contexts. Inversion was possible, but it was a general option (whose interpretive import is not always clear) with all narrow-scope indefinites (Stark 2006: 210–212) and also with universal quantifiers (Giusti 2010: 392–395).

For old Catalan, the study of algun by Bergareche and Saldanya (2011) highlights some interesting tendencies. First, Bergareche and Saldanya (2011: 225) give a characterization of the meaning of algun that comes close to the definition of epistemic indefinites: “[i]t represents a quantifier of the existential type, together with un, ‘a’, from which algun differed by having a greater level of indeterminacy: ‘at least one, one or more’ and never ‘just one’, which was the proper value of un”. An example corresponding to this definition is (29):

(29)

old Catalan (Jaume I, Fets: 187r, from Bergareche and Saldanya 2011: 226)

mester hi hauria que algú hi resposés
necessity:sg there have:3sg that someone:sg there answer:3sg

‘it would be necessary for someone to answer’

Second, the distribution in old Catalan confirms that the NPI use is invariantly transmitted from late Latin to all the varieties considered here: a 13th-century example of algun in the context of sentential negation (a possibility excluded in contemporary Catalan) is given in (30).

(30)

old Catalan (Usatges de Barcelona 157.18)

e negú per algun enguiny ni
and nobody:sg for any:sg trickery:sg nor
per art no sia gosat rebugar
for artifice:sg not be:3sg dared refuse

‘Let no one by any artifice or trickery dare to refuse to obey it (the judgment)’

Furthermore, Bergareche and Saldanya (2011: 227) observe that, if algun follows the noun, it is necessarily interpreted as an NPI, thus it can only occur in modal or negative contexts. A negative context featuring DP-internal inversion is exemplified in (31):

(31)

old Catalan (Exemples 6.31, from Bergareche and Saldanya 2011: 227)

Amich, no m’ ajusté ne m’ acosté a
friend not me join:1sg nor me approach:1sg to
neguna fembre, ne fiu mal ne desplaer a
no:sg woman:sg nor make:1sg evil:sg nor upset:sg to
home algun
man:sg any:sg

‘My friend, I did not join or approach any woman, nor did I cause any evil or upset to any man’

The existence of an intermediate stage characterized by DP-internal inversion in negative contexts persisting until a much later age is indirectly witnessed also by Par (1923: 46–47), who, in commenting on earlier language stages, observed that the postposed use of algun in negative contexts was still possible in his own variety. This confirms that DP-internal inversion becomes obligatory, at some point, in all those languages that retain both the epistemic and the NPI use (which is later lost in Catalan).

Equally in old Spanish and old Portuguese, as is to be expected given the situation in the contemporary languages, the descendants of aliquis maintain all uses attested in late Latin. Martins (2000) speaks of “polar versatility” with respect to the early Romance behavior of the continuations of aliquis and of other NPIs and provides ample exemplification.

Martins (2015b) contributes important observations on the distribution of DP-internal inversion in old Spanish and old Portuguese: she finds that, in both languages, inversion is possible, but it is not obligatory in negative contexts, as shown in (32) for old Portuguese.

(32)

old Portuguese (Demanda do Santo Graal 146 ed. Nunes, from Martins 2015b: 110)

ca nom vos pode ende vĩir algũũ bem
because not you:pl can:3sg from.that come any:sg good:sg

‘because this will not bring you any good’

Moreover, inversion was possible also in contexts with a positive interpretation, as we can see in (33), which forms a near-minimal pair with (32).

(33)

old Portuguese (Demanda do Santo Graal 64 ed. Nunes, from Martins 2015b: 109)

Se aqui ficardes em esta furesta, toste
If here stay:2pl in this:sg forest:sg soon
vos poderia vĩir ende mal algũũ
you:pl can:3sg come from.that harm:sg any:sg

‘If you stay here in this forest any longer, soon some harm may come to you.’

The example in (33) also evidences the ignorance component that characterizes epistemic indefinites: pending more accurate studies, this suggests that the positive use of old Portuguese is also a continuation of the (late) Latin one.

5 From Epistemic Indefinites to NPIs and NCIs

5.1 The Diachronic Cline

In the previous sections we have seen that the singular form of a lexical item whose etymological source had a positive meaning develops a negative use in all the Romance languages that continue it. The positive use can be more precisely characterized as epistemic and goes back to the Latin source. As for the negative use, which first emerges in late Latin, it is necessary to distinguish sensitivity to downward-entailing contexts (NPI use) from featural specification as an element of Negative Concord (NCI use), depending on the language.

Cases of determiners and pronouns changing from a “positive” to a “negative” meaning are well known from the literature (cf. Jäger 2010; Willis et al. 2013: 36–38 for a crosslinguistic overview). The best studied case concerns French personne ‘nobody’ and rien ‘nothing’, NCIs that find their origin in generic nouns (‘person’, ‘thing’), cf. Déprez (2000, 2011, Roberts and Roussou (2003: 146–154), Déprez and Martineau (2004), Roussou (2007), Labelle and Espinal (2014). Willis (2011), Willis et al. (2013: 28) label this recurrent change phenomenon the “Quantifier cycle” and connect it to Ladusaw’s (1993) “Jespersen Argument Cycle”, whose stages are schematically summarized in (34):

(34)
Quantifier Cycle/Jespersen Argument Cycle (Ladusaw 1993: 438)

“one thing” (plain indefinite) > “anything” (NPI) > “nothing” (NCI or Negative Indefinite)

The process involving aliquis and its Romance continuants can be considered an instantiation of the directional development in (34). However, it cannot be fully equated with other known cases with respect to its initial stage, since it does not fully square with the characterization of the “one thing” stage given by Willis et al. (2013: 27): “indefinites starting out as contextually unrestricted items”. As seen in §2.2, epistemic indefinites are pragmatically licensed by the interaction with (modal) operators in the clause, which allow their contribution to be meaningful in a pragmatic perspective, thus ensuring their felicity. Borrowing Kratzer and Shimoyama’s (2002) terminology, epistemic indefinites are selective with respect to the contexts in which they can occur, since they are semantically and pragmatically dependent on co-occurring operators (cf. Giannakidou 1998; Giannakidou and Quer 2013 for the notion of dependency).

Therefore, epistemic indefinites are not contextually unrestricted, but subject to a licensing relation. As discussed in Gianollo (2018: 93–94), the fact that epistemic indefinites are subject to a licensing relation, similarly to NPIs and NCIs, makes them likely candidates for diachronic processes of reanalysis, affecting the nature of the licensing relation itself. I argue that this is exactly what motivates the observed diachronic cline with aliquis and its Romance continuations, and in §5.2 I propose a way to model the reanalysis.

In concluding this section, note that, as a consequence of the specification of the initial stage for the change involving aliquis, the change’s general format takes a different shape: while Willis et al. (2013: 27–28, 36–38) model the process exclusively as a “contraction in the range of environments available for an item” (Willis et al. 2013: 36), we see that aliquis first undergoes a process of extension in the range of compatible environments, when the NPI uses emerge (in late Latin) and coexist with the epistemic uses. The contraction happens later, in some individual Romance languages, when NPI uses in the scope of sentential negation are reanalyzed as a form of syntactic dependency (Negative Concord). The first diachronic step consists in the overgeneralization of a pattern that, at some point, is perceived by the language users as systematic enough. This would be a case of innovation in which “the acquirer ‘boosts’ the regularity of a pattern in the input, ascribing it a formal status that it does not have in the adult grammar” (Biberauer 2019: 221).

5.2 The Role of DP-internal Inversion

For Ladusaw (1993), the process in (34) is crucially connected to the pragmatic properties of negation and interacts with the cycle of formal renewal that affects negative markers crosslinguistically (Jespersen’s Cycle): indefinites undergoing (34) are recruited into the system of negation as negation strengtheners in emphatic contexts (cf. also Kiparsky and Condoravdi 2006; Roussou 2007). In what follows, I build on this intuition and I motivate the late Latin extension of aliquis into negative contexts as due to the interaction between the epistemic indefinite and a scalar focus operator, which I take to be the determinant of emphatic strengthening, following Chierchia (2013). I furthermore argue that the interaction with the scalar focus operator is responsible for DP-internal inversion, which I take to be triggered by the determiner’s movement to a DP-internal Focus projection.

5.2.1 The Semantic Change

According to the characterization of epistemic indefinites proposed by Alonso-Ovalle and Menéndez-Benito (2010) and seen in §2.2, epistemic indefinites trigger a minimal widening of their domain of quantification: the quantificational domain must include more than one possible referential alternative. Epistemic indefinites are not used in the scope of negation because their contribution is vacuous: the pragmatic reasoning triggered by the minimal domain widening results in statements that are already implied by the assertion (Lauer 2012).

The effect is very different with NPIs, which are maximal domain wideners. With maximal domain widening, all possible alternatives, even the most peripheral ones, are considered (Kadmon and Landman 1993). Maximal domain widening removes contextual restrictions; hence, a negative sentence containing an NPI is true in a superset of contexts in which the same sentence with a plain existential would be true. This results in a strengthened assertion, and precisely this strengthening effect motivates the use of NPIs in a communicative perspective (cf. Israel 2011: 168–180 for a summary of strengthening-based approaches to NPIs). Thus, while epistemic indefinites typically weaken an assertion (with the purpose of avoiding an exhaustivity inference, or avoiding a false claim in the case of lacking evidence), NPIs result in a stronger assertion.

My proposal assumes that, when interacting with focus in emphatic negative constructions, aliquis and its Romance continuants become strengthened and behave as maximal domain wideners. The focus operator involved is what Krifka (2007) calls scalar focus, and Chierchia (2013) identifies as his E (= ‘even’) operator.

Focus evokes alternatives to the proposition in the scope of the focus operator. Scalar focus is considered to be responsible for emphatic negation readings, such as those emerging when minimizers or generalizers are used (Chierchia 2013 a.o.; Kiparsky and Condoravdi 2006; Roussou 2007). With scalar focus, alternatives are ordered along a contextually provided scale and the focused element represents an extreme value. Focus particles like even require scalar focus for their interpretation: in an example like (35), ‘a child’ is considered to be the least probable alternative among the individuals who could solve a maths problem. The alternative denotations are generated according to world knowledge and are ordered along a probability measure (a mathematician > a maths teacher > an adult…).

(35)
Even a child could solve this maths problem

When interacting with negation, scalar focus yields an emphatic assertion, whose communicative felicity is based on the fact that even the most probable alternative in the considered domain does not hold.[14] In (36) the fact that a drop may be spilled is the most probable alternative in case of a spilling event; negation reverses the entailments and the proposition “not a drop of this wine was spilled” denotes the least probable alternative, which is a suitable focus value for even.

(36)
Not even a drop of this wine was spilled

When scalar focus interacts with aliquis in the scope of negation, additional alternatives are evoked by the focus operator. These alternative sets undergo union with the set of alternatives in the epistemic indefinite’s quantificational domain. The result of this operation is maximal domain widening, which has the effect of strengthening the negative assertion.

Originally, the widening effect emerges contextually, through the interaction between focus, involved in emphatic negation, and the minimal domain widening triggered by aliquis. Diachronically, it becomes conventionalized in all the Romance continuations of aliquis that are used in negative contexts. In French and Italian, this is the only surviving use. In Catalan it gets lost, whereas in Spanish and Portuguese it coexists with a positive use. In §5.2.2 I analyze the obligatory inversion undergone by algún / algum in negative contexts as a result of the conventionalization of domain widening, that is, as a necessary interaction with a Focus operator in the syntax.

5.2.2 The Syntactic Implications

Based on the conclusions reached above, I propose that a semantically motivated syntactic operation derives DP-internal inversion between the descendants of aliquis and the noun in negative contexts. Martins (2015a, 2015b) treats inversion as a syntactic operation that builds an NPI in the syntactic component and is triggered by negative formal features; she assumes a DP-internal projection for negation (NegP) to which the determiner moves.

My proposal is similar in deriving inversion as a syntactic operation and in causally connecting it to polarity sensitivity. However, in my proposal the landing site is different: it is a Focus projection in the left periphery of the DP. My assumption is motivated by the semantic mechanism proposed in §5.2.1 and is compatible both with NPI uses (as found in Italian and Spanish) and with NCI ones (corresponding to the distribution in French and Portuguese).

The underlying DP structure I assume is presented in (37):

(37)
[DP [TopP [FocP [dP [NP ]]]]]

A rich articulation of the DP’s left periphery, comprising a projection for Topic and Focus, has been defended in a number of works (see Aboh et al. 2010 for an overview). For Romance, a parallelism between focus-related displacement in the DP and in the clause has been investigated by Bernstein (2001): I propose that DP-internal inversion with the descendants of aliquis belongs to the family of phenomena presented there. In (37), dP represents the projection at the top of the DP’s inflectional domain (Giusti 2006).

I propose that the quantificational determiner is first merged in dP. The semantically motivated association with focus triggers the determiner’s movement to the FocP projection. Inversion results from a further movement of the N(P) to the Topic position above the determiner’s landing site (cf. 38).

(38)
[DP [TopP [NP estudiante] [FocP alguno [dP alguno [NP estudiante]]]]]

N(P) movement to TopP is motivated pragmatically by the necessity to background the material in the restriction of the quantifier. This allows the postponed determiner to receive emphatic stress, prosodically aligning the focused element to the nuclear stress (in a way similar to what happens in Romance VOS orders, cf. the motivation of VOS orders in terms of the Nuclear Stress Rule in Zubizarreta 1998, and the movement to Topic of given information in Belletti’s (2001) analysis of Italian VOS orders).

In some cases only the head noun moves, possibly leaving additional DP-material in a position that linearly follows the determiner. Head movement of N0 is the only option available in Portuguese (cf. 39): Martins (2015b: 105, 113) attributes this to the incorporation operation she assumes between the determiner and the head noun, resulting in strict adjacency.

(39)

Portuguese (Martins 2015a: 404)

* Animal selvagem algum vive aqui
animal:sg wild:sg no:sg live:3sg here

‘No wild animal lives here’

Spanish (Martins 2015a: 413–414, 2015b: 114–115 fn. 14) and Italian (Gianollo 2018: 128–129) are different, allowing for intervening material between the noun and the determiner under certain conditions (cf. 40 for Italian). This suggests that in Spanish and Italian phrasal movement of the whole NP to the Topic position is possible.

(40)

Italian

Non abbiamo animale domestico alcuno in questa casa
not have:1pl animal:sg domestic:sg any:sg in this:sg house:sg

‘We (absolutely) have no domestic animal in this house’

If the hypothesis on DP-internal inversion explored above is tenable, it leads us to assume different lexical entries for the positive and the negative descendants of aliquis in those languages that continue both: the negative forms have to include a movement-triggering feature in their lexical entry, which renders DP-internal inversion obligatory. In French and Italian, which only continue the negative use, inversion remains optional: according to speakers’ judgments, in the contemporary varieties it is an emphatic variant confined to high registers. Also in the case of optional inversion, thus, scalar focus appears to be connected to the syntactic operation.

In Gianollo (2018: Ch. 3 and 5) I discuss how this proposal can be more precisely formalized in terms of formal focus features. Here I limit myself to remark that the assumption of a separate lexical entry for the negative variant is anyway necessary for Portuguese, where algum in DP-internal inversion constructions is an NCI. According to the framework I adopt (cf. §2.3), this means that it carries formal uninterpretable features for negation ([uNeg] in Zeijlstra 2004), which allow it to be self-licensing, that is, to trigger the insertion of a covert negative operator when it is not c-commanded by another overt negative element (e.g., when it is in subject position; cf. again Zeijlstra 2004). The assumption of uninterpretable formal features for negation is also necessary for the lexical entry of French aucun. In Spanish and Italian, by contrast, as in late Latin we have an NPI distribution, which in §2.3 I assumed to be linked to a form of semantic-pragmatic licensing.[15]

The change from NPI to NCI witnessed in French and Portuguese is a purely syntactic development, understood as the conventionalization of the licensing relation with negation in terms of syntactically active features (a form of grammaticalization). The crucial semantic change, which lies at the origin of this further development, is the extension to NPI contexts first observed in late Latin.

The conventionalization of the licensing relation is expected to lead to the loss of the emphatic effect connected to the originally focus-induced movement. Indeed, the results of my questionnaire (cf. fn. 2) show that while Spanish speakers find the use of inverted algún under negation emphatic, Portuguese speakers do not observe the same effect. This may be taken as an indication that the grammaticalization involved in the NPI > NCI shift is linked to the loss of the original pragmatic motivation for the movement. This mechanism reminds us of what happens with Jespersen’s Cycle when the original reinforcer of a negation marker replaces the older marker in the plain negation function, losing its emphatic value (cf. Gianollo 2018: 132–135 for further details).

6 Conclusions

This work attempts the study of a lexical item from the inventory of function words, which is diachronically very pertinacious and whose descendants in the various Romance languages display a remarkable combination of parallelism and differentiation.

I started this paper by presenting the variation in the descendants of Latin aliquis observed in the contemporary standard Romance languages. The attention focused in particular on epistemic uses in positive contexts and on NPI and NCI uses in the context of sentential negation. I modeled the space of variation as a set of conditions imposed by quantificational determiners on their quantificational domain and on the shape of their set of alternatives: epistemic determiners impose minimal domain widening, whereas NPIs impose maximal domain widening. The NCI use is derived from the NPI one as the grammaticalization of the licensing relation with the negative operator in terms of formal features.

The analysis of the initial stage for the process involving aliquis, that is, its characterization as epistemic indefinite in Classical Latin, allows us to pinpoint the origin of the interpretational dependency that requires aliquis to be in the scope of a suitable licensing operator. I proposed that this licensing relation is reanalyzed in late Latin, leading to the extension of aliquis into negative contexts.

The medieval Romance descendants of aliquis invariably continue both the positive (epistemic) and the negative (NPI) use. Further diachronic processes cause differentiation across Romance: some languages lose the epistemic use, some other languages lose the NPI one, and in some languages a reanalysis of the NPI use as an NCI syntactic dependency takes place. Spanish and Portuguese retain uses in both positive and negative contexts, but the latter are characterized by obligatory DP-internal inversion between the determiner and the noun. I connected this distribution to the semantic effect of focus on the quantificational domain, and I proposed a syntactic mechanism to account for obligatory DP-internal inversion.

The observed variation in the contemporary Romance languages finds its roots in the properties of the Latin ancestor and can be explained by an interesting case of semantic change, with repercussions on the syntax of nominal expressions. I hope to have shown the relevance of the diachronic study of elements of the functional lexicon for our understanding of the interaction between semantic and syntactic change. The diachronic development of aliquis and its descendants reveals interesting parallels between the Quantifier Cycle and Jespersen’s Cycle: both have in common the fact that, in negative contexts, emphatic focus interacts with the inferences triggered by nominal elements in its scope (quantificational determiners, minimizers, generalizers), yielding systematic meaning effects that may undergo grammaticalization during the diachronic development.


Corresponding author: Chiara Gianollo, Department of Classical Philology and Italian Studies, Università di Bologna, Bologna, Italy, E-mail

References

Aboh, Enoch, Norbert Corver, Marina Dyakonova & Marjo van Koppen. 2010. DP-internal information structure: Some introductory remarks. Lingua 120. 782–801. https://doi.org/10.1016/j.lingua.2009.02.010.10.1016/j.lingua.2009.02.010Search in Google Scholar

Alonso-Ovalle, Luis & Paula Menéndez-Benito. 2003. Some epistemic indefinites. Proceedings of NELS 33. 1–12.10.1093/acprof:oso/9780199665297.003.0001Search in Google Scholar

Alonso-Ovalle, Luis & Paula Menéndez-Benito. 2010. Modal indefinites. Natural Language Semantics 18. 1–31. https://doi.org/10.1007/s11050-009-9048-4.10.1007/s11050-009-9048-4Search in Google Scholar

Alonso-Ovalle, Luis & Paula Menéndez-Benito. 2011. Domain restrictions, modal implicatures and plurality: Spanish algunos. Journal of Semantics 28. 211–240. https://doi.org/10.1093/jos/ffq016.10.1093/jos/ffq016Search in Google Scholar

Alonso-Ovalle, Luis & Paula Menéndez-Benito (eds.). 2015. Epistemic indefinites. Exploring modality beyond the verbal domain. Oxford: Oxford University Press.10.1093/acprof:oso/9780199665297.001.0001Search in Google Scholar

Belletti, Adriana. 2001. ‘Inversion’ as focalization. In Aafke Hulk & Jean-Yves Pollock (eds.), Subject Inversion in Romance and the theory of Universal Grammar, 60–90. Oxford: Oxford University Press.Search in Google Scholar

Bergareche, Bruno Camus & Manuel Pérez Saldanya. 2011. From adjectives to quantifiers. Evidence from Old and Modern Catalan. Catalan Journal of Linguistics 10. 221–245.10.5565/rev/catjl.62Search in Google Scholar

Bernstein, Judy. 2001. Focusing the ‘right’ way in Romance determiner phrases. Probus 13. 1–29. https://doi.org/10.1515/prbs.13.1.1.10.1515/prbs.13.1.1Search in Google Scholar

Biberauer, Theresa. 2019. Children always go beyond the input: The Maximise Minimal Means perspective. Theoretical Linguistics 45(3–4). 211–224. https://doi.org/10.1515/tl-2019-0013.10.1515/tl-2019-0013Search in Google Scholar

Bortolussi, Bernard. 2015. Syntaxe des indéfinis latins. Quis, quisque, alius. Paris: Presses de l’Université Paris-Sorbonne.Search in Google Scholar

Chierchia, Gennaro. 2013. Logic in grammar: Polarity, free choice, and intervention. Oxford: Oxford University Press.10.1093/acprof:oso/9780199697977.001.0001Search in Google Scholar

Corblin, Francis. 1996. Multiple negation processing in natural language. Theoria 62(3). 214–259.10.1111/j.1755-2567.1996.tb00503.xSearch in Google Scholar

Crisma, Paola. 2012. Quantifiers in Italian. In Edward Keenan & Denis Paperno (eds.), Handbook of quantifiers in natural language, 467–534. Berlin: Springer.10.1007/978-94-007-2681-9_9Search in Google Scholar

Déprez, Viviane. 2000. Parallel (a)symmetries and the structure of negative expressions. Natural Language and Linguistic Theory 18. 253–342. https://doi.org/10.1023/a:1006449808181.10.1023/A:1006449808181Search in Google Scholar

Déprez, Viviane. 2011. Atoms of negation: An outside-in micro-parametric approach to negative concord. In Richard Ingham & Pierre Larrivée (eds.), The evolution of negation. Beyond the Jespersen cycle, 221–272. Berlin/New York: de Gruyter.10.1515/9783110238617.221Search in Google Scholar

Déprez, Viviane & France Martineau. 2004. Micro-parametric variation and Negative Concord. In Julie Auger, J. Clancy Clements & Barbara Vance (eds.), Contemporary approaches to Romance linguistics: Selected papers from the 33rd LSRL, 139–158. Amsterdam: Benjamins.10.1075/cilt.258.08depSearch in Google Scholar

Devine, Andrew M. & Laurence D. Stephens. 2013. Semantics for Latin. An introduction. Oxford: Oxford University Press.Search in Google Scholar

Etxeberria, Urtzi & Anastasia Giannakidou. 2017. A unified analysis of Spanish algún/algunos: anti- specificity, plurality, and NP-anaphora. Ms.Search in Google Scholar

Fălăuş, Anamaria. 2007. Le paradoxe de la double négation dans une langue à concordance négative stricte. In Franck, Floricic (ed.), La négation dans les langues romanes, 75–97. Amsterdam: Benjamins.10.1075/lis.26.05falSearch in Google Scholar

Fălăuş, Anamaria. 2014. (Partially) free choice of alternatives. Linguistics and Philosophy 37. 121–173.10.1007/s10988-014-9146-8Search in Google Scholar

Farkas, Donka. 2002. Extreme non-specificity in Romanian. In Claire Beyssade, Reineke Bok-Bennema, Frank Drijkoningen & Paola Monachesi (eds.), Romance languages and linguistic theory 2000. 127–151. Amsterdam: Benjamins.10.1075/cilt.232.09farSearch in Google Scholar

Foulet, Lucien. 1919. ‘Quelque’. Romania 45. 220–249. https://doi.org/10.3406/roma.1919.5158.10.3406/roma.1919.5158Search in Google Scholar

Foulet, Lucien. 1970. Petite syntaxe de l'ancien français, 3rd revised edn. Paris: Champion.Search in Google Scholar

Giannakidou, Anastasia. 1998. Polarity sensitivity as (non)veridical dependency. Amsterdam: Benjamins.10.1075/la.23Search in Google Scholar

Giannakidou, Anastasia & Josep Quer. 2013. Exhaustive and non-exhaustive variation with free choice and referential vagueness: Evidence from Greek, Catalan, and Spanish. Lingua 126. 120–149. https://doi.org/10.1016/j.lingua.2012.12.005.10.1016/j.lingua.2012.12.005Search in Google Scholar

Giannakidou, Anastasia & Hedde Zeijlstra. 2017. The Landscape of Negative Dependencies: Negative Concord and N-Words. In Martin Everaert & Henk van Riemsdijk (eds.), The Wiley Blackwell Companion to Syntax, 2nd edn. London: Wiley Blackwell.10.1002/9781118358733.wbsyncom102Search in Google Scholar

Gianollo, Chiara. 2018. Indefinites between Latin and Romance. Oxford: Oxford University Press.10.1093/oso/9780198812661.001.0001Search in Google Scholar

Giusti, Giuliana. 2006. Parallels in clausal and nominal periphery. In Mara Frascarelli (ed.), Phases of interpretation, 163–184. Berlin: Mouton de Gruyter.10.1515/9783110197723.3.163Search in Google Scholar

Giusti, Giuliana. 2010. Le espressioni di quantità. In Giampaolo Salvi & Lorenzo Renzi (eds.), Grammatica dell’italiano antico. Bologna: Il Mulino.Search in Google Scholar

Haspelmath, Martin. 1997. Indefinite pronouns. Oxford: Oxford University Press.Search in Google Scholar

Ingham, Richard. 2011. Grammar change in Anglo-Norman and Continental French. The replacement of non-affirmative indefinite nul by aucun. Diachronica 28(4). 441–467. https://doi.org/10.1075/dia.28.4.01ing.10.1075/dia.28.4.01ingSearch in Google Scholar

Ingham, Richard & Amel Kallel. 2014. Evidence from a correspondence corpus for diachronic change in French indefinites 1450-1715. In Maj-Britt Mosegaard Hansen & Jacqueline Visconti (eds.), The diachrony of negation, 213–234. Amsterdam: Benjamins.10.1075/slcs.160.08ingSearch in Google Scholar

Iordăchioaia, Gianina & Frank Richter. 2015. Negative concord with polyadic quantifiers. The case of Romanian. Natural Language and Linguistic Theory 33. 607–658.10.1007/s11049-014-9261-9Search in Google Scholar

Israel, Michael. 2011. The grammar of polarity. Pragmatics, sensitivity, and the logic of scales. Cambridge: Cambridge University Press.10.1017/CBO9780511975288Search in Google Scholar

Jayez, Jacques & Lucia Tovena. 2006. Epistemic determiners. Journal of Semantics 23(3). 217–250. https://doi.org/10.1093/jos/ffl002.10.1093/jos/ffl002Search in Google Scholar

Jayez, Jacques & Lucia Tovena. 2011. The meaning and (a bit of) the history of quelque. In Lucia Tovena (ed.), French determiners in and across time, 111–139. London: College Publications.Search in Google Scholar

Jäger, Agnes. 2010. Anything is nothing is something. On the diachrony of polarity types of indefinites. Natural Language and Linguistic Theory 28(4). 787–822. https://doi.org/10.1007/s11049-010-9113-1.10.1007/s11049-010-9113-1Search in Google Scholar

Kadmon, Nirit & Fred Landman. 1993. Any. Linguistics and Philosophy 16(4). 353–422. https://doi.org/10.1007/bf00985272.10.1007/BF00985272Search in Google Scholar

Kiparsky, Paul & Cleo Condoravdi. 2006. Tracking Jespersen’s Cycle. In Brian Joseph & Angela Ralli (eds.), Proceedings of the Second International Conference of Modern Greek dialects and linguistic theory, 172–197. Patras: University of Patras.Search in Google Scholar

Kratzer, Angelika & Junko Shimoyama. 2002. Indeterminate pronouns: The view from Japanese. Proceedings of the 3rd Tokyo Conference on Psycholinguistics, 1–25.10.1007/978-3-319-10106-4_7Search in Google Scholar

Krifka, Manfred. 2007. Basic notions of information structure. Working Papers of the SFB632, Interdisciplinary Studies on Information Structure (ISIS) 6. 13–56. https://publishup.uni-potsdam.de/opus4-ubp/frontdoor/deliver/index/docId/1430/file/ISIS06.pdf.Search in Google Scholar

König, Ekkehard. 1991. The meaning of focus particles: A comparative perspective. London: Routledge.Search in Google Scholar

Labelle, Marie & M. Teresa Espinal. 2014. Diachronic changes in negative expressions: The case of French. Lingua 145. 194–225. https://doi.org/10.1016/j.lingua.2014.04.002.10.1016/j.lingua.2014.04.002Search in Google Scholar

Ladusaw, William. 1992. Expressing negation. In Chris Barker & David Dowty (eds.), Proceedings of SALT, vol. 2, 237–259: Ohio State University: Ohio State Working Papers in Linguistics 40.10.3765/salt.v2i0.3030Search in Google Scholar

Ladusaw, William. 1993. Negation, indefinites, and the Jespersen cycle. In Joshua S. Guenter, Barbara A. Kaiser & Cheryl C. Zoll (eds.), Proceedings of the 19th annual meeting of the Berkeley Linguistics Society, 437–446. Berkeley: Berkeley Linguistics Society.10.3765/bls.v19i1.1515Search in Google Scholar

Lauer, Sven. 2012. Some news about irgend- and algún. Ms: Stanford University.Search in Google Scholar

Ledgeway, Adam. 2009. Grammatica diacronica del napoletano. Tübingen: Niemeyer.10.1515/9783484971288Search in Google Scholar

Leonetti, Manuel. 2004. Specificity and Differential Object Marking in Spanish. Catalan Journal of Linguistics 3. 75–114. https://doi.org/10.5565/rev/catjl.106.10.5565/rev/catjl.106Search in Google Scholar

López, Luis. 2012. Indefinite objects. Scrambling, choice functions, and differential marking. Cambridge, MA: MIT Press.10.7551/mitpress/9165.001.0001Search in Google Scholar

Marouzeau, Jules. 1922. L’ordre des mots dans la phrase latine, vol. 1 Les groupes nominaux. Paris: Champion.Search in Google Scholar

Martí, Luisa. 2008. The semantics of plural indefinite noun phrases in Spanish and Portuguese. Natural Language Semantics 16. 1–38. https://doi.org/10.1007/s11050-007-9023-x.10.1007/s11050-007-9023-xSearch in Google Scholar

Martins, Ana Maria. 2000. Polarity Items in Romance: Underspecification and Lexical Change. In Susan Pintzuk, George, Tsoulas & Anthony Warner (eds.), Diachronic syntax. Models and mechanisms, 191–219. Oxford: Oxford University Press.Search in Google Scholar

Martins, Ana Maria. 2015a. Ordem de palavras e polaridade. Inversão nominal negativa com algum / alguno e nenhum. Diacrítica 29(1). 399–426. http://www.scielo.mec.pt/pdf/dia/v29n1/v29n1a16.pdf.Search in Google Scholar

Martins, Ana Maria. 2015b. Negation and NPI composition inside DP. In Theresa Biberauer & George Walkden (eds.), Syntax over time, 102–121. Oxford: Oxford University Press.10.1093/acprof:oso/9780199687923.003.0007Search in Google Scholar

Mauri, Caterina, Silvia Ballarè, Eugenio Goria, Massimo Cerruti & Francesco Suriano. 2019. KIParla corpus: A new resource for spoken Italian. In Raffaella Bernardi, Roberto Navigli & Giovanni Semeraro (eds.), Proceedings of the 6th Italian conference on computational linguistics CLiC-it.Search in Google Scholar

Milner, Jean-Claude. 1979. Le système de la néǵtion en français et l'opacité du sujet. Langue française 44. 80–105. https://doi.org/10.3406/lfr.1979.6171.10.3406/lfr.1979.6171Search in Google Scholar

Par, Anfos. 1923. Sintaxi catalana segons los escrits en prosa de Bernat Metge. Halle: Niemeyer.10.1515/9783112324929Search in Google Scholar

Penka, Doris. 2011. Negative indefinites. Oxford: Oxford University Press.10.1093/acprof:oso/9780199567263.001.0001Search in Google Scholar

REW = Wilhelm Meyer-Lübke. 1935. Romanisches etymologisches Wörterbuch 3. vollständig neubearbeitete Auflage. Available online urn:nbn:de:bvb:355-ubr07799–0. Heidelberg: Winter.Search in Google Scholar

Roberts, Ian & Anna Roussou. 2003. Syntactic change. A minimalist approach to grammaticalization. Cambridge: Cambridge University Press.10.1017/CBO9780511486326Search in Google Scholar

Roussou, Anna. 2007. Minimalism and diachronic syntax: The development of negative expressions. In Eleni Agathopoulou, Maria Dimitrakopoulou & Despina Papadopoulou (eds.), Selected papers on theoretical and applied linguistics: Festschrift for Prof. E. Panagopoulos, vol. 17 1, 11–27. Thessaloniki: Monochromia.Search in Google Scholar

Schwarzschild, Roger. 2002. Singleton Indefinites. Journal of Semantics 19. 289–314. https://doi.org/10.1093/jos/19.3.289.10.1093/jos/19.3.289Search in Google Scholar

Stark, Elisabeth. 2002. Indefiniteness and specificity in Old Italian texts. Journal of Semantics 19(3). 315–332. https://doi.org/10.1093/jos/19.3.315.10.1093/jos/19.3.315Search in Google Scholar

Stark, Elisabeth. 2006. Indefinitheit und Textkohärenz: Entstehung und Semantische Strukturierung Indefiniter Nominaldetermination im Altitalienischen. Tübingen: Niemeyer.10.1515/9783110925838Search in Google Scholar

Traina, Alfonso & Tullio Bertotti. 1985. Sintassi normativa della lingua latina. 3rd edn. Bologna: Cappelli.Search in Google Scholar

Willis, David. 2011. Negative polarity and the Quantifier Cycle: comparative diachronic perspectives from European languages. In Pierre Larrivée & Richard Ingham (eds.), The evolution of negation. Beyond the Jespersen cycle, 285–323. Berlin: Mouton de Gruyter.10.1515/9783110238617.285Search in Google Scholar

Willis, David, Christopher Lucas & Anne Breitbarth. 2013. Comparing diachronies of negation. In David Willis, Christopher Lucas & Anne Breitbarth (eds.), The history of negation in the languages of Europe and the Mediterranean, vol. 1: Case studies, 1–50. Oxford: Oxford University Press.10.1093/acprof:oso/9780199602537.003.0001Search in Google Scholar

Zamparelli, Roberto. 2007. On singular existential quantifiers in Italian. In Ileana Comorovski & Klaus von Heusinger (eds.), Existence: Semantics and syntax, 293–328. Berlin: Springer.10.1007/978-1-4020-6197-4_11Search in Google Scholar

Zeijlstra, Hedde. 2004. Sentential negation and negative concord. Amsterdam: Universiteit van Amsterdam dissertation.Search in Google Scholar

Zeijlstra, Hedde. 2016. Negation and negative dependencies. Annual Review of Linguistics 2. 233–254. https://doi.org/10.1146/annurev-linguistics-030514-125126.10.1146/annurev-linguistics-030514-125126Search in Google Scholar

Zubizarreta, Maria Luisa. 1998. Prosody, focus, and word order. Cambridge, MA: MIT Press.Search in Google Scholar

Published Online: 2020-09-22
Published in Print: 2020-11-18

© 2020 Chiara Gianollo, published by De Gruyter, Berlin/Boston

This work is licensed under the Creative Commons Attribution 4.0 International License.

Downloaded on 1.2.2023 from https://www.degruyter.com/document/doi/10.1515/probus-2020-0005/html
Scroll Up Arrow