Navigation – Plan du site
Perspectives diachroniques

Negation and indefinites in Late Latin

La négation et les indéfinis en latin tardif
Chiara Gianollo
p. 277-286

Résumés

Dans cette étude, j’examine l’interaction entre la négation phrastique et les indéfinis négatifs dans des textes latins tardifs (à partir du iiie siècle av. J.-C.), avec le but de repérer, à cette époque, des traces des développements qu’on peut retrouver par la suite dans les langues romanes anciennes. Je propose que la persistance de l’ordre Objet-Verbe avec les indéfinis négatifs du latin tardif est un signe d’une restructuration précoce dans le système de la négation phrastique, qui prélude aux systèmes des langues romanes. Je présente une interprétation économique de ce processus diachronique, que j’interprète comme un changement crucial des traits formels du marqueur de négation nōn, changement qui a eu des conséquences significatives pour son rapport avec les pronoms indéfinis.

Haut de page

Texte intégral

1. Introduction

1The diachronic process known as Jespersen’s Cycle involves the lexical substitution of a language’s marker of sentential negation. The new element typically originates as an optional strengthener (e.g. French pas), and a doubling stage frequently precedes the demise of the old marker. Recent comparative investigation of Jespersen’s Cycle has unveiled its profound effects on the whole system of negation, beyond the form of the negative marker itself: a process of change affecting the negative marker typically affects the behavior of negative indefinites as well (cf. Willis et al., 2013, for a cross-linguistic overview).

2The workings of Jespersen’s Cycle are well known from the history of Early Romance languages, and especially from the history of French. But, in fact, the earliest Latin documents show that such a cycle had just been completed, taking to the grammaticalization of nōn as reinforced form of the original Indo-European negative particle . This process has been comprehensively treated in recent work (Fruyt, 2008; 2011). My aim here is to discuss some novel evidence pointing to a further development related to Jespersen’s Cycle in Late Latin, which paves the way towards the Romance systems.

  • 1 For reasons of space I used abbreviations for some technical notions introduced in the text. I list (...)

3To this end, I will first present Late Latin data concerning the distribution of negative indefinites in object function (Section 2). I will then argue that their consistent preverbal placement may be interpreted as a signal for an underlying change in the syntactic status of the negative marker, and may explain why negative indefinites decay and are mostly not continued in the Early Romance languages. To this end, I will need to briefly introduce the main points in the interpretation of negation systems and Jespersen’s Cycle offered by recent generative models (Section 3). I will then come back to the analysis of the Late Latin data (Section 4) and summarize my conclusions in Section 5.1

2. The position of negative objects in Late Latin

  • 2 Cf. Polo, 2004; Salvi, 2005; Spevak, 2005; Sznajder and Bortolussi, 2010.

4The shift from O(bject)V(erb) to VO in the history of Latin has been extensively investigated from different points of view. Here I will not attempt to summarize the findings in this domain, referring the reader to the overview of the literature in Ledgeway (2012, p. 225-35). It is well known that already Early and Classical Latin displayed a pragmatically and stylistically conditioned alternation between preverbal and postverbal objects. Moreover, the placement of the object per se cannot be treated independently from the syntactic behavior of the inflected verb, which shows an increase of non-final orders already in the Classical grammar (a fact which may in principle be due to an independent syntactic phenomenon, cf. Danckaert, 2012; 2014). Late Latin texts display much variation, with some genres being more conservative than others, so that it would be obviously wrong to treat Late Latin as a whole as a “VO language”. In-depth studies on individual texts have shown that the postverbal position of objects is subject to precise pragmatic factors.2

5I would like to argue here that, despite this very complex situation, it is nonetheless possible to single out a phenomenon that seems to yield quite clear-cut data, at least in the texts that I analyzed, namely, the position of negative objects. Despite the growing tendency to position the object postverbally, accusative forms of the negative indefinite pronouns and adjectives with object function regularly appear preverbally. This fact is noted by Spevak (2005, p. 248) for the Itinerarium Egeriae, where however very few cases are found. But the result is replicated in many coeval texts, and also in earlier ones, starting in the 1st cent. BC. I show the results of my corpus study in the two Tables below, respectively for Classical (Table 1) and Late Latin (Table 2).

  • 3 I excluded elliptical constructions and embedded infinitival clauses, the latter since they quite i (...)

6I run my queries over selected texts contained in the LLT-A; sometimes I only checked for one or two indefinites of the negative series, but for some authors I have data for nemo, nihil, nullus (“all acc.” in the Tables; null* indicates that accusative forms for all genders have been checked). Under “Tot(al)” I counted all accusative forms for the respective indefinite; under “Rel(evant)” I only considered those which occur in a context where the position of the negative indefinite object with respect to the verb can be safely and significantly assessed.3

Table 1: Classical Latin – negative indefinites in object function

Author / Text

Form

Tot/Rel

OV

VO

other

Plautus

neminem

26/14

6

7

1

Terence

neminem

10/6

2

4

Cicero, Epist.

neminem

65/34

20

13

1

Varro

all acc.

15/8

6

0

2

Vitruvius

all acc.

11/6

5

0

1

Livy

neminem

85/31

26

1

4

Celsus

null* + neminem

18/5

5

0

Petronius

all acc.

47/29

28

0

1

7Table 1 shows that, while in Plautus, Terence, and Cicero VO orders are found quite normally with negative indefinites, OV order is predominant or categorical in authors of treatises, like Varro and Vitruvius, and most distinctly in Livy, whose work supplies us with more extensive evidence.

Table 2: Late Latin – negative indefinites in object function

Author / Text

Form

Tot/Rel

OV

VO

other

Pass. Perp.

all acc.

3/2

2

0

Itin. Eg.

null*

2/2

2

0

Aug. Serm.

neminem

64/48

46

0

2

Vulgata

null* + neminem

62/42

39

3

Vulgata – Evang.

nihil

25/22

19

3

Oros. Hist.

all acc.

51/30

30

0

Greg. T. Franc.

null*

43/27

27

0

8The comparison between Table 1 and Table 2 shows that the tendency towards OV orders with negative indefinites becomes regularity in the Late Latin of the 3rd-4th cent. AD (and is confirmed also in a later author like Gregory of Tours). In fact, OV order with negative indefinites is even more consistent in Late Latin texts than in early Classical Latin texts, where their flexible placement conforms to the general pattern observed for the Classical grammatical system. The steady OV order for negative indefinites does not seem to be paralleled by similar phenomena affecting negative polarity items (henceforth, NPIs), e.g. quisquam, or other quantificational elements, e.g. omnis, cf. Table 3.

Table 3: Comparison with NPIs and quantifiers

Author / Text

Form

Tot/Rel

OV

VO

other

Aug. Serm.

quemquam

17/13

6

3

4

Aug. Serm.

omnem

100/88

10

37

41

Vulgata

quemquam

25/15

2

13

Vulgata – NT

omnem

68/37

10

27

  • 4 For Germanic languages, cf. Jónsson, 1996; Svenonius, 2000; Pintzuk and Taylor, 2006; for Romance l (...)

9The diachronic persistence of OV orders with negative objects during the shift from OV to VO is well known from the history of other languages, notably in the Germanic and Romance family.4 For Latin, I propose that the consistency of OV orders is due to the new syntactic role that negative indefinites come to play in the system of negation in Late Latin. I will thus argue for the existence of a specific syntactic operation targeting them (and not other objects) because of their negative value, and forcing them to appear in a precise area of the clause. In order to motivate my proposal, I will introduce my theoretical assumptions in Section 3 and come back to the analysis of the data in Section 4.

3. Negation systems in synchrony and in diachrony

10I follow recent theoretical literature in assuming three basic types of negation systems:

  • a. Double Negation (DN): each negatively marked element conveys a semantic negation (Latin, German);

  • b. Strict Negative Concord (NC): the negatively marked indefinites (n-words) always co-occur with a sentential negative marker and convey only one semantic negation (French, Romanian, Russian, Modern Greek);

  • c. Non-strict Negative Concord: the negatively marked indefinites (n-words) co-occur with a negative marker, conveying one semantic negation, if they linearly follow the finite verb; however, if preceding the finite verb they cannot co-occur with a negative marker in a single negation reading (Italian, Spanish).

  • 5 See Molinelli, 1988; Orlandini, 2001; Bertocchi et al., 2010.

11Latin’s very rich system of indefinites sharply distinguishes between negative indefinites (1) and NPIs (2).5 Negative indefinites always bring about a negative reading, i.e. they may not co-occur with another overt expression of negation without yielding a ‘double negation’ reading (3). In this respect Latin patterns with e.g. Modern Germanic languages like German or English.

(1)

Neminem reperies qui neget.
‘You will not find anyone who would deny it.’
(Cic. Verr. 2, 2, 152)

(2)

Non ante tibi ullus placebit locus.
‘Before that (otherwise) no place will please you.’ (Sen. Epist. 28, 2)

(3)

Aperte enim adulantem nemo non videt.
‘No one does not recognize someone who is blatantly flattering.’ (Cic. Lael. 99)

12All (early) Romance languages, instead, exhibit Negative Concord: a negative marker can co-occur with a negative indefinite, yielding only one semantic negation (4).

(4)

Non troverai nessuno che possa negarlo.
‘You will not find anyone who would deny it.’

  • 6 Zanuttini, 2010; Parry, 2013; Poletto, forthc.

13Romance languages exhibiting non-strict NC are particularly relevant for my discussion since they show an interesting asymmetry between two areas of the clause, separated by what is known in the generative literature as Infl(ection) / T(ense) Phrase (see Ledgeway, 2012, p. 119-50 for this category in Latin and Romance). The finite verb in Romance is realized in this position, and the standard preverbal negative markers derived from Latin nōn (e.g. It. non, Sp. and Cat. no, Pt. não, Fr. ne, Rom. nu) occupy a position above it (NegP-1 in Zanuttini, 1997). Recent studies on Early Romance have confirmed that this has been the case since the beginning of attestation.6

14Note that in non-strict NC systems the possibility of co-occurrence of the indefinite with a negative marker is dependent on the surface configuration (more precisely, on the configuration reflecting the hierarchical relations between elements in their final landing sites), and not on the underlying grammatical structure (original position of insertion, i.e. merge): both subject and object indefinites cannot co-occur with the negative marker if pre-Infl (5), and must co-occur with it if post-Infl (6).

(5)

a. Nessuno studente ha consegnato il compito.
‘No student handed in the homework.’
b.
NIENTE ha mangiato!
‘S/he ate NOTHING!’ (capitals indicate emphatic focus)

(6)

a. Non ha consegnato il compito nessuno studente.
‘No student handed in the homework.’
b. Non ha mangiato niente.
‘S/he did not eat anything.’

15We have to conclude that indefinites in non-strict NC languages like It. nessuno, niente have an ambiguous status, since their ability to negate autonomously depends on their position: for this reason they are called “n-words”, to distinguish them from unambiguously negative indefinites. This ambiguity has been captured in various ways. In Table 4 I present the analysis proposed by Zeijlstra (2004; 2008), which is based on differences in the featural composition of the various elements in negation systems.

Table 4: Negation systems (Zeijlstra, 2004; 2008)

Type

Negative Marker

Indefinites

Double Negation

[Neg]

[Neg] (Negative Indefinites)

strict Negative Concord

[uNeg]

[uNeg] (n-words)

non-strict Negative Concord

[iNeg]

[uNeg] (n-words)

16Negative indefinites of DN systems are distinguished from n-words of NC systems in that the former are intrinsically semantically negative. In DN systems, thus, the non-co-occurrence of negatively marked elements is captured by assigning to all elements a self-licensed semantic feature [Neg] that does not trigger (licensing-related) syntactic phenomena after its insertion.

17In NC systems, instead, n-words are only formally marked for negation, but this feature is uninterpretable [uNeg] and needs to enter a syntactic relation (called “Agree”) with an interpretable [iNeg] feature. In the simplified system adopted here (but see Zeijlstra 2004; 2008 for the complex picture), I will assume that the [iNeg] feature is always hosted in the NegP-1 projection. If the n-word is below (i.e. c-commanded by) the position NegP-1 where the [iNeg] feature sits (see Figure 1), the [iNeg] feature can and must be realized by the negative marker (adverb or particle), which enters an Agree relation with the [uNeg] on the n-word.

Figure 1: Position of NegP-1

Figure 1: Position of NegP-1

18If, instead, for some independent reasons (raising to subject position, information structural requirements) the n-word is above the position where [iNeg] sits, i.e. before the finite verb, in non-strict NC systems it suffices alone to convey sentential negation and it does not co-occur with the negative marker. I will assume that in this case the n-word is local enough to the NegP-1 projection to satisfy its requirements (see Gianollo, 2016, for more details).

19This analysis allows for a number of typological and diachronic predictions. According to Zeijlstra (2004), since in DN languages negation is purely semantic, DN languages lack a formal, syntactically active feature for negation [i/uNeg]: as a consequence, they do not grammaticalize a Neg projection, as there are no phenomena (e.g. the establishment of Agree relations) that could provide sufficient evidence for it during acquisition. From this the following prediction arises:

20Negative heads are predicted not to be available in non-Negative-Concord languages. There is no language without Negative Concord that exhibits a negative marker that is a syntactic head (Zeijlstra, 2004, 2008).

  • 7 Note that this formulation is probably too strong: we may well imagine languages with a negative he (...)

21Having a negative head means having an element that projects its own category NegP. This category requires establishing visible relations (i.e. Negative Concord) with other negatively marked elements, if these are present in the clause.7 In languages without NC, instead, each negative element contributes a logical operator for negation: the negative marker attaches as adjunct phrase to a verbal projection, with no NegP in the clausal spine.

  • 8 See Rowlett, 1998; van Kemenade, 2000; van Gelderen, 2004; Jäger 2008.

22We thus see that there is a further dimension distinguishing negation systems cross-linguistically: whether a negative marker has the status of a head (X0) or of a phrase (XP). If phrasal, the negative marker may either occupy the Specifier of NegP, or, as adjunct, the outer Specifier of another category. This connects in interesting ways with the generative interpretation of Jespersen’s Cycle (schematically summarized in (7), but see van der Auwera, 2009, for its actual complexity). Syntactically, this process may be seen as the reinforcement of an X0 negative marker by means of an XP element, which may subsequently become the “real” negator.8 In course of time, the cycle may proceed further, leading to the reanalysis of the XP element as an X0, and triggering a further reinforcement process.

(7)

Stage I: simple negative marker – X0 (Old French ne)
Stage II: reinforced negative marker – X
0 + XP complex (French ne … pas)
Stage III: renewed simple negative marker - XP and later X
0 status (Colloquial French pas)

23If we adopt a global perspective on negation systems, it emerges quite clearly that changes affecting the negative marker must have an effect on the indefinites interacting with it: in particular, a negative marker changing to head status, i.e. becoming a [i/uNeg] head of a functional projection NegP, will cease to be compatible with [Neg] negative indefinites in a single-negation reading, since both elements always bring about a logical negative operator (cf. Table 4). Thus, it is expected that reaching Stage III of Jespersen’s Cycle will cause a chain-reaction in the system of indefinites interacting with negation. In the next section I argue that such a situation holds in Late Latin.

4. Late Latin as a ‘concealed Negative Concord’ language

  • 9 Cf. Fruyt, 2008; 2011, p. 708-23.

24The Classical Latin negative marker nōn is clearly the product (Stage III) of a pre-historic but plausibly relatively recent Jespersen’s Cycle involving the original negation < IE *ne and the numeral ‘one’ used as a scale-evoking minimizer (8).9 Thus, a XP status seems probable for nōn in light of its origin as a reinforcer (and of further structural phenomena, on which see Gianollo, 2016).

(8)

nōn < + *oinom = oenum ( = ūnum) ‘not (even) one’

25As for the position of nōn, an important generalization stems from standard descriptions of Latin: nōn regularly precedes the finite verb, i.e. in analytical forms (9) it appears immediately before the auxiliary, not before the participle, cf. Kühner and Stegmann (1912, p. 818).

(9)

Romanus equitatus ipsum quidem regem Elatiae adsecutus non est.
‘But the Roman cavalry did not reach the king of Elatea himself.’ (Liv. 36, 19, 10)

  • 10 Cf. Danckaert, 2012; 2014; Devine and Stephens, 2006.

26Given a number of assumptions on Latin clause structure,10 this represents unambiguous evidence for the fact that the negative marker precedes the category Infl. In other words, the position of Latin negation is the same as that of the continuations of nōn in Romance, as discussed in Section 3.

27In Late Latin nōn stays in the same pre-Infl position; however, I argue that nōn is reanalyzed as a head (probably already in Late Classical Latin). Such reanalysis leads to the completion of a round of Jespersen’s Cycle and, in fact, to the conditions to start a new one, i.e. to a new Stage I in (7). The reanalysis of a specifier XP as a head is a diachronically frequent phenomenon, which obeys the structure-minimizing tendency known as ‘Spec-to-Head principle / Head Preference Principle’ (cf. van Gelderen, 2004).

28I further assume that the feature on Neg0 becomes formal interpretable [iNeg]. This feature is syntactically active in all negative clauses: this means that it may establish a syntactic relation with other elements within its scope containing a negative feature, yielding thus Negative Concord structures. But Latin negative indefinites have a [Neg] feature: they thus bring about a logical negation of their own and are incompatible with the new X0 negative marker in the single negation reading. We thus reach the scenario depicted at the end of Section 3; in what follows I would like to propose, however, that there is a way to escape such incompatibility.

  • 11 I refer here to the well-known examples of ‘emphatic’ negation often discussed in the literature (c (...)

29Recall that in non-strict NC languages n-words preceding the inflected verb (pre-Infl) suffice to express sentential negation and may not co-occur with the negative marker. Thus, in the pre-Infl area the surface behavior of non-strict NC and DN languages overlaps, despite the different featural composition of the indefinite items. According to my proposal, the prerequisites for NC (mainly, a negative marker at Stage I of a new Jespersen’s Cycle) are already present in Late Latin. Why, then, do we not see NC? Why does the co-occurrence of the negative marker with the negatively marked indefinites in a single-negation reading remain at the low rates we know from Classical Latin (cf. Molinelli, 1988)?11 I argue that this is linked to the fact that negative objects may still precede the inflected verb, and thus also the NegP projection. In this area of the clause, they superficially behave like n-words of non-strict NC systems: they do not co-occur with the negative marker and negate the clause by themselves.

30As seen in Section 2, negative indefinites systematically surface in pre-verbal position in the Late Latin texts I analyzed, even more systematically so than in Classical Latin, despite a general shift towards a VO grammar. Thanks to this configuration, negative indefinites escape the c-command domain of the new NegP projection and surface above it, without co-occurring with an overt negative marker, exactly as pre-Infl n-words in the Romance languages. According to this analysis, Late Latin may be considered a “concealed Negative Concord language”: the prerequisites for NC are transmitted to Romance, representing a case of syntactic pertinacity and shared inheritance.

  • 12 Only nullus is productively continued in Early Romance. For nemo, cf. Rom. n-word nimeni ‘no one’.

31Note that, under this interpretation, nemo, nihil, nullus, etc. are not reanalyzed in their featural composition, but remain [Neg] in Late Latin. In Gianollo (2016) I argue that this is a welcome result in light of the extensive lexical replacement taking place in this domain during the transmission to Romance: a lack of reanalysis predicts, in fact, the disappearance of the “old” negative indefinites, which cease to be compatible with the new NC system.12

5. Summary and conclusions

32In my analysis, the main change event affecting the syntax of negation from Latin to Romance is situated in Late Latin and concerns the featural specification (and the co-occurring change in phrasal status) of the negative marker nōn. As a consequence, negative indefinites become licit only when preverbal. This preludes to their demise in Romance, where an array of new indefinites interacting with negation is grammaticalized.

33In the process of change towards Romance we also observe some very pertinacious characteristics, like the position of the negative marker in the clause. More generally, what looks like a major typological change (from a DN to a NC system) reduces to a discrete reanalysis of the featural specification of a functional lexical item (nōn).

Haut de page

Bibliographie

Auwera, J. van der, 2009, The Jespersen Cycles, in E. van Gelderen (ed.), Cyclical Change, Amsterdam, p. 35-71.

Bertocchi, A., Maraldi, M., and Orlandini, A., 2010, Quantification, in Ph. Baldi and P. Cuzzolin (eds.), New Perspectives on Historical Latin Syntax, vol. 3, Berlin / New York, p. 19-173.

Danckaert, L., 2012, Latin Embedded Clauses, Amsterdam.

Danckaert, L., 2014, The derivation of Classical Latin Aux-final clauses. Implications for the internal structure of the verb phrase, in K. Lahousse and S. Marzo (eds.), Romance Languages and Linguistic Theory 2012, Amsterdam, p. 141-160.

Devine, A. M. and Stephens, L. D., 2006, Latin Word Order. Structured Meaning and Information. Oxford.

Ernout, A. and Thomas, F., 1953, Syntaxe latine, Paris.

Fruyt, M., 2008, Négation et grammaticalisation en latin, in Revue de Linguistique Latine du Centre Alfred Ernout De lingua Latina 1.

Fruyt, M., 2011, Grammaticalization in Latin, in Ph. Baldi and P. Cuzzolin (eds.), New Perspectives on Historical Latin Syntax, vol. 4, Berlin / New York, p. 661-864.

Garzonio, J. and Poletto, C., 2012, On niente: Optional negative concord in Old Italian, Linguistische Berichte 230, p. 131-153.

Gelderen, E. van, 2004, Economy, innovation, and prescriptivism: From Spec to Head and Head to Head, Journal of Comparative Germanic Linguistics 7, p. 59-98.

Gianollo, C., 2016, Indefinites between Latin and Romance, Habilitationsschrift, University of Cologne.

Jäger, A., 2008, History of German Negation, Amsterdam.

Jónsson, J. G., 1996, Clausal Architecture and Case in Icelandic, Amherst.

Kayne, R. S., 1975, French Syntax: The Transformational Cycle, Cambridge MA.

Kemenade, A. van, 2000, Jespersen’s cycle revisited. Formal properties of grammaticalization, in S. Pintzuk, G. Tsoulas, and A. Warner (eds.), Diachronic Syntax. Models and Mechanisms, Oxford, p. 51-74.

Kühner, R. and Stegmann, K., 1912, Ausführliche Grammatik der lateinischen Sprache, vol. 1, Hannover.

LLT-A: Library of Latin Texts. Series A, Turnhout, http://www.brepolis.net.

Ledgeway, A., 2012, From Latin to Romance: Morphosyntactic Typology and Change, Oxford.

Molinelli, P., 1988, Fenomeni della negazione dal latino all’italiano, Florence.

Orlandini, A., 2001, Négation et argumentation en latin, Louvain.

Parry, M., 2013, Negation in the history of Italo-Romance, in D. Willis, C. Lucas, and A. Breitbarth (eds.), The History of Negation in the Languages of Europe and the Mediterranean, vol. 1: Case studies, Oxford, p. 77-118.

Pintzuk, S., and Taylor, A., 2006, The loss of OV order in the history of English, in A. van Kemenade and B. Los (eds.), The Handbook of the History of English, Oxford, p. 249-278.

Poletto, C., forthcoming, Negation, in A. Ledgeway and M. Maiden (eds.), The Oxford Guide to the Romance Languages, Oxford.

Polo, Ch., 2004, Word Order between Morphology and Syntax, Padua.

Rowlett, P., 1998, Sentential Negation in French, Oxford.

Salvi, G., 2005, Some firm points on Latin word order: The left periphery, in K. É. Kiss (ed.), Universal Grammar in the Reconstruction of Ancient Languages, Berlin / New York, p. 429-456.

Spevak, O., 2005, Itinerarium Egeriae. L’ordre des constituants obligatoires, Mnemosyne 58, p. 235-261.

Svenonius, P., 2000, Quantifier movement in Icelandic, in P. Svenonius (ed.), The Derivation of VO and OV, Amsterdam, p. 255-292.

Sznajder, L. and Bortolussi, B., 2010, Ordres VSO et SVO dans la Vulgate, Journal of Latin Linguistics 11 (1), p. 273-300.

Willis, D., Lucas, C., and Breitbarth, A., 2013, Comparing diachronies of negation, in D. Willis, C. Lucas, and A. Breitbarth (eds.), The History of Negation in the Languages of Europe and the Mediterranean, vol. 1: Case studies, Oxford, p. 1-50.

Zanuttini, R., 1997, Negation and Clausal Structure: A Comparative Study of Romance Languages, Oxford.

Zanuttini, R., 2010, La negazione, in G. Salvi and L. Renzi (eds.), Grammatica dell’italiano antico, vol. 1, Bologna, p. 569-582.

Zeijlstra, H., 2004, Sentential Negation and Negative Concord, dissertation, University of Amsterdam.

Zeijlstra, H., 2008, Negative Concord is syntactic agreement, Ms. University of Amsterdam, http://ling.auf.net/lingbuzz/000645.

Haut de page

Notes

1 For reasons of space I used abbreviations for some technical notions introduced in the text. I list them here for convenience: pre-/post-Infl (before / after the position of the finite inflected verb = Infl); NPI: negative polarity item; DN: Double Negation systems; NC: Negative Concord systems; NegP: negative projection.

2 Cf. Polo, 2004; Salvi, 2005; Spevak, 2005; Sznajder and Bortolussi, 2010.

3 I excluded elliptical constructions and embedded infinitival clauses, the latter since they quite invariantly show (S)OV orders and might have biased the results towards my conclusion. The column “other” includes accusative subjects of certain verbs (e.g. paenitet, fallit) and complements of prepositions.

4 For Germanic languages, cf. Jónsson, 1996; Svenonius, 2000; Pintzuk and Taylor, 2006; for Romance languages, cf. Kayne, 1975; Garzonio and Poletto, 2012.

5 See Molinelli, 1988; Orlandini, 2001; Bertocchi et al., 2010.

6 Zanuttini, 2010; Parry, 2013; Poletto, forthc.

7 Note that this formulation is probably too strong: we may well imagine languages with a negative head and no negatively marked indefinites, i.e. no n-words. In these languages the indefinites interacting with negation may be non-negatively marked NPIs. In Gianollo (2016) I argue that this was in fact the case for pre-historical Latin.

8 See Rowlett, 1998; van Kemenade, 2000; van Gelderen, 2004; Jäger 2008.

9 Cf. Fruyt, 2008; 2011, p. 708-23.

10 Cf. Danckaert, 2012; 2014; Devine and Stephens, 2006.

11 I refer here to the well-known examples of ‘emphatic’ negation often discussed in the literature (cf. Ernout and Thomas, 1953, p. 154-5), but indeed quite infrequent until very late texts (where one may imagine that they are influenced by the developing Romance systems).

12 Only nullus is productively continued in Early Romance. For nemo, cf. Rom. n-word nimeni ‘no one’.

Haut de page

Table des illustrations

Titre Figure 1: Position of NegP-1
URL http://pallas.revues.org/docannexe/image/3757/img-1.png
Fichier image/png, 9,5k
Haut de page

Pour citer cet article

Référence papier

Chiara Gianollo, « Negation and indefinites in Late Latin », Pallas, 102 | 2016, 277-286.

Référence électronique

Chiara Gianollo, « Negation and indefinites in Late Latin », Pallas [En ligne], 102 | 2016, mis en ligne le 20 décembre 2016, consulté le 22 octobre 2017. URL : http://pallas.revues.org/3757 ; DOI : 10.4000/pallas.3757

Haut de page

Auteur

Chiara Gianollo

Researcher and Lecturer, Università di Bologna
chiara.gianollo@unibo.it

Haut de page

Droits d’auteur

Licence Creative Commons
Pallas – Revue d'études antiques est mis à disposition selon les termes de la licence Creative Commons Attribution - Pas d'Utilisation Commerciale - Pas de Modification 4.0 International.

Haut de page
  • Logo Presses universitaires du Midi
  • Revues.org