Conventional implicature and language change: the romanian paradigm of identity
A evolução cíclica dos pronomes românicos de identidade (ex.: însuș i e același ) fornece dados interessantes para documentar o impacto de fatores pragmáticos na mudança lingü.stica, uma vez que ela mostra como palavras que trazem implicaturas convencionais semelhantes podem se tornar alternativas adequadas para substituir itens homófonos.
The evolution of Lat. sibi “REFL-DAT” in Romanian provides interesting data on the cliticization of a full word and its transformation into a bound morpheme.1 It also shows how homonymic clashes can contribute to the loss of an entire morphemic paradigm and how, thanks to shared pragmatic functions, new alternates may acquire the possibility of carrying the same conventional implicature as the lost items. The development in question can be summarized as follows:
Stage I. In Old Romanian (--16th c.-- 18th c.) și, the reflex of Lat. sibi “REFL-DAT”, becomes an enclitic bound morpheme expressing “coreferentiality”. Attached to personal pronouns and to deictic expressions (pronouns and adverbials), -și served to develop an entire paradigm of means denying an expected non-coreferentiality. After quantifiers and indefinite pronouns it became even a mere intensifier.
Stage II (18thc.--): The bound morpheme -și is restricted to a reduced number of combinations, namely the emphatic pronoun însuși “self’”, the pronoun of identity același “same”, and the temporal adverbials acuși (now.self), atunceși (then.self) “right away”. It is very likely that this limited distribution is due mainly to the ambiguity created by its occurrence in similar contexts to its homophonous possessive dative. The pronoun însuși (a compound of the preposition în, întru “in(to)” + *isu (cf. Lat. ipsu “self, same, he”) + -și), becomes the standard expression of “self ”.
Stage III (19th c. --): The bound morpheme –și is no longer productive. The emphatic pronoun însuși “self ” becomes unpopular due to its highly irregular morphology and syntactic constraints. Consequently, other means of expressing an unexpected identity such as singur “alone” or chiar “(it is) clear (that P)” tend to replace it thanks to their shared pragmatic values of denial.
1. Markers of an unexpected identity
1.1 The bound morpheme –și and the emphatic pronouns
In Old Romanian (16thc. -18thc.) și (< Lat. sibi “REFL-DAT”), with its graphic variants ș, șu˘, și, became an enclitic marker of coreferentiality. The Latin reflexive dative sibi was characterized by two features that could predict its further evolution, namely: (i) as a reflexive, it marked the fact that the same referent was assigned two Roles in the given event (one Role being encoded as the syntactic subject in most cases) and (ii) as a dative, it referred mainly to the Experiencer / Beneficiary and was pragmatically exploited in various ways. According to Ernout & Thomas (1972: 184), the Latin reflexive still preserves its etymological meaning of “oneself ” (cf. Fr. “soi-même”). Flobert (1975: 387-388) defines the Latin reflexive pronoun as expressions of a high degree of voluntarism (“une volonté délibérée de soi sur soi”). Thanks to this value it can co-occur even with the middle (i.e. the forms in –r, so-called deponents, middle or passive forms), which also presupposes the fact that the referent of the subject is both the doer and the undergoer (in Flobert’s words, the form in -r represents a “dédoublement du sujet tout à la fois agissant et agi” – see (1)).
Generally speaking, the dative case on its own carries a special pragmatic connotation, since it points to the most salient constituent after the Agent. As Hymann & Zimmer (1976: 189-212) have pointed out, dative is more marked as to topicalization and focalization processes than any other oblique case. In Givón’s (1984: 139-141) topicalization hierarchy, the semantic role Dative (usually expressed by the dative morphological case) comes immediately after the Agent as the most likely candidate for the topic when the Agent is not specified. This pragmatic connotation must have also contributed to the reinterpretation of the reflexive -și as a preferred means for reinforcing the emphatic value of the Old Romanian personal and demonstrative pronouns, first in the dative and then in the accusative.
In Old Romanian (16th c.) the bound morpheme –și (cf. Lat. sibi “self: DAT”) could be attached to all kinds of pronouns: personal (luiș “of/to himself ”, loruși “of/to themselves”, mineși “myself ”, tineși “yourself ”, noiși “ourselves”, voiși “yourselves”), demonstratives (același “that.self ”, i.e. “same”), and deictic adverbs (e.g. acuși “now.self ”, i.e. “immediately”, etc.). All these forms co-occur synchronically in the Romanian older texts, so the stages of the spread of –și as a marker of coreferentiality have to be reconstructed. Within the framework of the extended model of prototypical semantics, it is possible to reconstruct diachronic semantic links on the basis of the synchronic relations between synonymous lexemes (see Kleiber, 1990: 180-181; Geeraerts, 1987)[4,5]. In our opinion, the only possible reasonable scenario for the evolution of Lat. sibi in Romanian should present the following sequencing:
First of all, -și (ș, șu˘) occurred after the dative lui (cf. Vulg.Lat. *illui – Cl. Lat. illi “to him”) to express coreferentiality with the subject in contexts where lui alone could be ambiguous. As (2) shows, after a noun, lui could refer to either (i) two coreferential arguments of the same predicate or (ii) two coreferential arguments of different predicates.
In (2) lui2 refers to an indirect object previously identified (the prodigal son) and not to the subject of deade “gave” [the landlord]. But lui1 (in satulu lui “his village”) could be decoded as being coreferential with the subject of tremease “he sent” only thanks to the knowledge of the information provided by the whole co-text of the parable. The model for the spread of -și from dative to possessives must have been offered by the stressed personal pronoun in the dative (lui “of/to him”), which could be either an indirect object or an attribute. This double function of lui has been favored by the persistence of the Latin construction called dativus adnominalis in Old Romanian (see (3)); cf. Fr. fils à papa lit. “son to papa”, i.e. “daddy’s boy”.
In (4) the addition of –și clearly disambiguates the “possessive” lui. The compound luișu refers unambiguously to the same person as the subject:
But, as shown by (5), the reflexive possessive adjective său “his-MASC.SG” (cf. Lat. suus) has in fact the same function of signaling co-referentiality:
The difference between luiși and the reflexive adjective său rests on the pragmatic level: as it will be demonstrated below, the compounds with –și (including luiși) deny an expected non-coreferentiality, whereas the reflexive adjective său does not.
Secondly, the dative compounds offered the model for the spread of –și as a marker of coreferentiality to the accusative of the 3rd person (eluși “himself-ACC” (see (7)), eiși “themselves-ACC” (Înv: 508), and even to other persons: mineși “myself- ACC” (Înv: 509), tineși “yourself-ACC”, noiș “ourselves-ACC” (CÎ:18) — see Densusianu 1961.II: 118-119). The fact that a dative was preferred over the accusative for attributing the pragmatic function of emphasis to personal pronouns is therefore due to both a syntactic factor and its pragmatic functions.
The pronominal compounds with -și may refer to a prominent topical constituent that is not the subject of the same clause. In (6), for example, luișu marks the unexpected coreferentiality with the topical Experiencer – i “him-DAT” (functioning as an indirect object):
The compound forms with -și co-occur frequently with the reflexive pronoun in order to confirm the identity in question, as an emphatic pronoun:
According to the maxim of quantity, the shorter utterance (8) would have conveyed the same idea of stabbing, but without the implicature that “according to the ethical principles of the community in question, one is not supposed to do so”.
As I hope to have demonstrated elsewhere (see Manoliu 1994: 192-194), the emphatic pronouns carry the conventional implicature that denies an expectation that the predicate in question applies to non-coreferential arguments. In symbolic logic terms, the expectation denied in (7) may be formulated as follows:
EXP7: ∃x ∃y (V(x, y)) . ~ (x ≡ y),
where ∃ is the existential quantifier “there is a…”; x would represent the first argument of the predicate “to stab”, the Agent; y represents the second argument of the predicate “to stab”, the Patient, whereas the symbol ~ represents the denial “it is not true that”. In other words, there is an Agent x and there is a Patient y, and the predicate “kills” applies to x, but it is not true that x is co-referential with y, so the predicate does not apply to two arguments which refer to the same person as both Agent and Patient simultaneously. The meaning asserted by (7) is:
S7: ∃x ∃y (V(x, y)) . ~ (~ (x ≡ y)),
i.e. “there is an x (the Agent) and a y (the Patient) and it is not true that the predicate ‘to stab’ does not apply to x as both the Agent and the Patient at the same time.” In brief, the expressions of “self ” carry the conventional implicature that denies the fact that the predicate applies to two non-coreferential arguments. So it forcefully confirms the coreferentiality of two arguments of the same predicate. In our opinion, this is the reason behind the label ‘pronouns of reinforcement’. In (9), the stressed reflexive accusative sineși (= sine “him/herself ” + -și) doubles the first reflexive se (direct object), as a strongly emphatic pronoun:
In (10), -și is added even to the possessive adjective săi “his/her/ their”; compare (5) and (10):
1.2 -Și and the identity pronoun “same”
With demonstratives, -și has a different function, namely it serves to express the meaning of “same”. In order to explain this change, it is necessary to account for the difference between “same” and “self ” in pragmatic terms. Let us thus analyze the utterance (11)2:
The expectation of (11) is that one does not sin and repent so quickly. In brief,
EXP11: ∃x “he” ∃y “hour1” (Vsin (x, y)) . ∃x ∃z “hour2” (Vrepent (x, z)) . ~ (y ≡ z).
The asserted meaning of (11) is thus:
S11: ∃x “he” ∃y “hour1” (Vsin (x,y)) . ∃x ∃z “hour2” (Vrepent (x, z)) . ~ (~ (y ≡ z)).
In other words, același “same” denies the expectation that the referent provided by the world of common beliefs for the argument y of the first predicate (“sin”) is not identical with the argument z of the second predicate (repent) and asserts that they are identical.
The difference between “self ” and “same” may be thus expressed in pragmatic terms as follows: “self ” confirms ‘the identity between two arguments of the same predicate’, whereas “same” confirms ‘the identity between the arguments of different predicates’3.
The reinterpretation of the compound of the distal demonstrative + -și as “same” was favored by the following features of the two components. The value of ‘coreferentiality’ brought in by –și is conjugated with the focus on the ‘novelty of the referent and/ or of the predication’ expressed by the demonstrative. As Kleiber (1992: 623), for example, emphasizes:
In brief, the demonstratives are strong signals of inviting the addressee to identify the referent as a new entity or as an already known entity to which a new predicate applies. In other words, in the presence of a demonstrative, the new predicate may apply either to (a) an argument coreferential with an argument of a previous predicate or (b) a new referent (when the demonstrative is used as an indexical). Such a context is incompatible with the idea of ‘coreferential arguments of the same predicate’. It is then explicable why the addition of –și (confirming coreferentiality) to a demonstrative (focusing on the novelty of the predicate) will activate meaning (a) and will result in the interpretation of the whole compound as the expression of “same”, which, as already shown, confirms the fact that the argument of one predicate is coreferential with the argument of another predicate. It is perhaps interesting to recall at this point the fact that Lat. ipse “self ” was also reinterpreted as “same” when co-occurring with demonstratives (see 2.2.1 below). The invariant pragmatic function shared by both “self ” and “same” may be defined as “the denial of an expected non-coreferentiality”.
When co-occurring with the proximity demonstrative (as in acestași), -și is just an additional marker intensifying the cataphoric value of acest “this”:
When -ș(u˘) is combined with temporal deictic adverbials, the confirmation of ‘identity between two moments’ is reinterpreted as ‘immediateness’ (short span of time between successive events): e.g. acmușu˘ “now.REFL” (13); atunceș “then.REFL” (14).v
The spatial deictic aci “here” combined with –și ends up by also expressing “immediateness”.
Combined with iară “again” as in iarăși “again (and again)”, -și reinforces the meaning of ‘repetition’, because iară alone was on its way of becoming a weak adversative conjunction (comp. (10) above and (16) below).
1.3. –Și as an intensifier
When -și is attached to other classes of constituents with no anaphoric function, its pragmatic interpretation as a marker of confirmation is converted into “increase in assertiveness”. In brief, the confirmation marker -și becomes an intensifier.
(i) As such, -și may follow an indefinite/interrogative pronoun: cineși “whoever”, cinreși, cinrescuși “whoever”; cinevași “somebody”, careși “each [of them]”, oareși-care “any [one]”.
(ii) It may also follow an indefinite quantifier: integral: totuluș “whole”; partitive: cîtuși [de puţin]“however [little]”, or an ordinal numeral: întîiași/dintîiași “the very first [time], from the beginning”.
2. Stage II (--18 th c.): the decay of the bound morpheme – ș i
During the period in question the bound morpheme -și becomes restricted to a reduced number of combinations. This reduction is probably due to the following factors:
a. homonymy with its semi-cliticized reflexive variant ș(˘u/i) expressing coreferentiality with the subject, as an indirect object of either the “beneficiary” or the “whole/possessor” (with both alienable and inalienable possessions). The contexts in which the reflexive possessive –ș(i/u˘) occurred after nouns constituted a favorable position in which confusions between the possessive and the emphatic -și could arise (see (23) below)
b. other expressions acquire the conventional implicatures carried by the compounds of –și, and hence competition between forms (for example, the pronoun însuși “himself ”, as an alternate noun modifier, the adjectival singur “alone”, the adverbial: chiar “even”, etc.).
Let us examine even briefly these concurrent factors.
2.1 SIBI proper: Rom. reflexive dative pronouns
The reflexes of the reflexive dative pronoun Lat. sibi, namely și, (ș, șu,˘), are first and foremost clitics that could be attached to all kinds of words ending in a vowel. In (20), for example, it is attached to the verb and has the function of an indirect object co-referential with the subject:
The use of the semi-cliticized reflexive variant as the ‘dative of the Whole/Possessor/Beneficiary’ constitutes another factor that undermined the use of the bound morpheme -și as a means of reinforcing the pragmatic functions of personal pronouns and deictics. As in other Romance languages, the reflexive of the “whole” could be used in conjunction with the accusative of the parts of the body4:
But the dative reflexive can also be used for the Beneficiary/ Possessor of an alienable possession even when not in contact with the human body:
In no Romanian text does the emphatic –și (șu) occur after a noun, because, on the one hand, the noun does not express co-referentiality by itself and, on the other, it would be in competition with its reflexive homophone expressing Beneficiary/Possessor, as shown by (23) and (24).
Moreover the reflexive indirect object și/u˘ could also be attached to the subject personal pronoun elu˘ “he”. The contracted form elu-șu˘ then becomes homophonous with the emphatic pronoun in the accusative. Compare (25) below and (7) above:
2.2. Lat. IPE - Rom. însuși, as an alternate reinforcement pronoun
2.2.1 Lat. IPSE
According to Ernout & Thomas (1972:189), ipse “est proprement un intensif, qui s’emploie avec une idée d’opposition latente” (is an intensive proper that is used with an idea of latent opposition). In other words, as any emphatic pronoun, ipse may be defined as a signal of “unexpected coreferentiality of two arguments of the same predicate”5. Let us consider the following example:
In (27) the use of ipsos (instead of the mere reflexive accusative se) in the ‘accusative + infinitive’ construction implies that the ambassadors might have suspected that somebody did not want to let them deliver their message in person. An even more interesting example of the role of ipse in denying an expected non-coreferentiality is provided by (28), where ipsae co-occurs with a reflexive pronoun:
The fact that the doors opened by themselves may not have been considered as an usual phenomenon in the everyday Roman life. According to the maxim of quantity, if such an event would have met the common beliefs, the utterance valvae se aperuerunt “the doors opened” would have been the normal choice. But in the given cultural context, characterized by the common belief that an external force has to act for opening doors, ipse is a sign of denying the expectation provided by the shared knowledge of a historically determined linguistic community. When combined with other demonstrative pronouns (hic ipse, iste ipse, ille ipse), ipse is virtually synonymous with idem (according to Ernout & Thomas (1972: 191) in such contexts “ipse se rapprochait de idem”). In V. Lat. ipse alone could also carry the conversational implicature of idem “same” as shown by the following utterance:
In Vulgar Latin, ipse started to loose its pragmatic value of “confirming an unexpected coreferentiality”, as shown by its co-occurrence with other “identity markers” such as -met: e.g. egomet ipse “I.and.nobody.else” + “self ” or metipse “self.self ”; cf. the resulting forms in Romance languages: Fr. même “self, same, even”, Sp. mismo “same”, Pg. mesmo, It. medesimo “same, self ”.
In spoken Latin ipse could replace other demonstratives such as iste or hic as shown by the corresponding Romance demonstratives: O. Sp. eje, Occ. eis, O.Pg. eiso, Sp. ese, Pg. esse “this-2nd” Aromanian nîs, năs, Istroromanian ăns “this” (Pușcariu 1975 s.v. 870). In some areas it then became a personal pronoun (cf. It. essi “they”), then a focalizer pointing to a salient constituent (see (30)) and even a definite article (cf. Sard. su “the”)6
2.2.2 The Romanian emphatic pronoun însuși
As a consequence of the loss of its illocutionary force of confirming an unexpected identity, the Romanian pronoun însu (deriving from ?in + ipsu) developed into a mere personal pronoun whose anaphoric function was reinforced by the addition of the article –l (< Lat. ille) as in însul (see (31)) and the compound dînsul (de “from”+ însul (see (32)). However its counterpart originating in the demonstrative ille “that” took over its functions as in most of the Romance languages, as shown by the fact that already in Old Romanian, însu had a very limited distribution. It could occur mainly in combination with a preposition: într-însu “in it-MASC”(CV: 248); într-însa “in it.the-FEM” (Ureche: 94); pre însul “on him.the” (Ureche:95); de înse “of them-FEM.PL”(Ureche:83), dentr-însa “from it.the-FEM.SG”(Ureche:121); cf. Cont. Rom. într-însul “in it/ him.the-MASC” and într-însa “in it/her.the-FEM.7
The reflexive –și was a welcome addition to the pronoun însu for expressing the confirmation of “an unexpected coreferentiality of the arguments of the same predicate”. In Old Romanian, the emphatic însuși alone could function as a pronoun (see 33), which is unacceptable in modern Romanian (see the corresponding cont. Rom. expression in (34)):
In (35) însușu˘ follows a demonstrative subject:
As an adjective it could also precede its head noun (36).
It could also double a personal pronoun:
As has been pointed above, in Old Romanian texts, însuși seems to have been the only possible candidate after nouns, since the enclitic –ș(u/i˘) had the function of a possessive dative (see (23) and (24) above). From nouns, însuși must have spread to pronouns, to eliminate also the possible confusion with the contracted form elu-șu˘ “he + to.himself ” (see (25) above).
3. Stage III: Modern Romanian
3.1. A new paradigm of identity
The rich paradigm of markers of a denied expected non-coreferentiality was reduced to the following combinations, which are also current in contemporary Romanian: același “same” (38) and însuși “self ’” (39, 40):
3.2 Stage IV (18th c. --): The decay of the emphatic pronouns
If același is still the sole form for “(the) same”(but see Manoliu 1987:421-424 for the tendency for it to be replaced by tot “also” + acela “that”), însuși “self ” has a similar fate as its precursors. It becomes vulnerable due to two factors: (a) morphological complexity: it has a highly irregular inflexion which has no match in any other nominal paradigm: (i) its gender and number are marked by a change in the stem vowel: -u/ă/i/e-; (ii) its gender agreement is governed by the gender of the referent in the 1st and 2nd persons and by the head noun in the 3rd person; (iii) moreover, the stress falls on the first syllable and, consequently, the complex final markers are unstressed (see the table of its inflexion in (41)) and (b) syntactic restrictions (for example, însuși cannot occur after a noun in the genitive (see (42)).
As shown by (42), when determining a noun, însuși is replaced by a complex construction namely: reflexive possessive (său/sa “his/ her”) + the adjective propriu “own”.
Confusions between these forms occur rather frequently in both Old and contemporary Romanian: see O. Rom. (43), where the 3rd person replaces însumi, and (44), where it replaces însuţi (in Gheţie1997: 127), or Cont. Rom.(45), where the singular form replaces the plural înseși (see Iordan et al.1967:133).
Consequently însuși tends to be replaced by other expressions capable of denying an expectation of non-identity, such as singur “alone”, “he and nobody else” (
3.3 Adjectival SINGUR “alone”
The core meaning of singur (cf. Lat. singulus) carries a conventional implicature denying the expectation that an additional participant could be involved in the event, confirming the fact that the given participant and nobody else should be considered for the argument in question. Such a pragmatic value is close enough to the one carried by însuși, which confirms the fact that the predicate applies to two coreferential arguments, excluding any other argument. It is thus explicable that in contexts such as (46) – (48), singur is synonymous with însuși. As such, singur may be found already in old texts as a subject emphatic pronoun and as a modifier of either a personal pronoun or a noun:
(i) as a subject pronoun:
(ii) following a personal pronoun:
(iii) preceding a noun:
In contemporary Romanian, when it functions as a subject, singur is preferred to însuși.
In (50) singur replaces însuși after the personal subject pronoun ea:
3.2.2. The adverbial chiar for însuși
Chiar originates in the adverbial use of the adjective (cf. Lat. clarum “clear”), with the meaning “(it is) clear (that P)”, as shown by its values in Old Romanian texts: “exactly, clearly, precisely, indeed, truly” (see Densusianu, 1961.2: 165) as well as by some of its contemporary contextual values (see (51)). It has a similar pragmatic function of an emphatic pronoun in the sense that it serves as a marker confirming the truth-value of an utterance referring to an unexpected quality or event.
In (51), chiar serves to deny the expectation that “he is not that stupid”. The replacement of însuși by chiar is explicable in pragmatic terms, since, in combination with an NP, they may carry a similar conventional implicature that confirms the fact that the predicate applies to an unexpected candidate for the referent of the modified constituent. As examples (52) and (53) show, according to our common beliefs, the “queen” is an unlikely candidate for the argument of the verb “to meet in a coffee shop in Davis”8:
The theoretical interest of the history of the Romanian emphatic pronouns is two-fold:
1. The cyclic evolution of the emphatic pronouns in Romanian shows how homonymic clashes contribute to the loss of an entire morphemic paradigm and how, thanks to shared pragmatic features (means of denial, confirmation of coreferentiality), new analytical expressions may replace the old forms that become less appropriate for carrying the implicature in question.
2. In agreement with the extended model of prototypical semantics, the split evolution of the reflexive dative pronouns provides interesting evidence for diachronic semantic reconstruction on the basis of attested synchronic variants.
Aeth: Siluiae uel potius Aetheriae Peregrinatio ad loca sancta, ed. by W. nd
Cicero, Diu: Cicero, Marcus Tullius. De divinatione, De fato, Timaeus, ed. by Remo Giomini, [Bibliotheca scriptorum Graecorum et Romanorum Teubneriana 46]. Leipzig: Teubner, 1975.
CÎ: Diaconul Coresi. Carte cu învăţătură (1581), ed. by Sextil Pușcariu and Alexie Procopovici, Bucharest: Socec. & Co., 1914.
Costin: Miron Costin. Opere alese. Letopiseţul Ţărîi Moldovei. De neamul moldovenilor, Viaţa lumii [16th c.], ed by Liviu Onu, Bucharest: Editura Știinţifică, 1967.
CP: Camil Petrescu. Opere, 3 , ed. by Al. Rosetti and Liviu Călin. Bucharest: Minerva, 1981.
CT: Tetraevanghelul diaconului Coresi (15611), ed. by Timuș Piteșteanu. Bucharest, 1889.
Curtius: Q. Curti Rufi Historiarum Alexandri Magni Macedonis libri qui supersunt, iterum recensuit Edmundus Hedicke. Editio maior. Leipzig: Teubner, 1908.
CV: Codicele Voroneţean [15th-16th cs.], ed. by Mariana Costinescu. Bucharest: Minerva, 1981.
Frag. Tod.: Fragmentul Todorescu, in Gheţie, Ion et al., eds. Texte românești din secolul al XVI-lea, Bucharest: Editura Academiei,1982.
GA: VP: Gabriela Adameșteanu, Vară, primăvară. Bucharest: Cartea Românească, 1989.
GA: DE: Gabriela Adameșteanu, Drumul egal al fiecărei zile. Bucharest: Litera, 1992.
Înv: Învăţături preste toate zilele (1642), 1-2, ed. by W. van Eeden, Amsterdam: Rodopi, 1985.
Moxa: Moxa, Mihail, Cronica universală (17thc.), ed. by G. Mihăilă. Bucharest: Minerva, 1989.
Nec: Ion Neculce. Opere. Letopiseţul Ţării Moldovei și O samă de cuvinte [18th c.], ed. by Gabriel Ștrempel. Bucharest: Minerva, 1982
Plaut, Ps: Plautus, Titus Maccius. Pseudolus, edited by M.M. Wilcock. Bristol: Bristol Classical Press, 1987.
Ureche: Grigore Ureche, Letopiseţul Ţării Moldovei [16th c], ed. by Liviu Onu. Bucharest: Editura Știinţifică, 1967.
- L’adjectif demonstratif dans la langue de la Bible latine. Abel Fritz. Tübingen: Niemeyer; 1971.
- Istoria Limbii române. 1. Originile, 2. Secolul al XVI-lea Densusianu Ovidiu. Bucharest: Editura Știinţifică; 1961.
- “Pragmatique linguistique” Ducrot Oswald. In: Parret Herman, ed. Le langage en contexte. Etudes philosophiques et linguistiques de pragmatique. Amsterdam: Benjamins; 1980 .
- “Great Expectations: an intensive self-analysis” Edmonson J. A, Plank F. Linguistics and Philosophy.1978;2:373-413.
- Syntaxe latine Ernout Alfred , Thomas François . Paris: Klincksieck; 1972.
- “Demonstrative Pronouns and Definite Article in Latin and the Romance languages” Faingold Eduardo D. Papiere zur Linguistik.1996;54(1):67-82.
- Les verbes déponents latins des origines à Charlemagne Flobert David. Paris: Les Belles Lettres; 1975.
- “Cognitive Grammar and the History of Lexical Grammar” Geerarerts D. In: Rudzka B, ed. Topics in Cognitive Grammar. Amsterdam: John Benjamins; 1987 .
- Istoria limbii române literare. Epoca veche (1532-1780) Gheţie Ion et al. Bucharest: Editura Academiei Române; 1997.
- Syntax. A Functional Typological Introduction Givón T. Amsterdam & Philadelphia: John Benjamins; 1984.
- “Embeded topic in French”, Subject and Topic Hymann L., Zimmer K. E.. New York, San Francisco & London: Academic Press; 1976.
- Structura morfologică a limbii române contemporane Iordan Iorgu, Guţu-Romalo V, Niculescu Al.. Bucharest: Editura Știinţifică; 1967.
- La sémantique du prototype Kleiber Georges. Paris: P.U.F.; 1990.
- “Anaphore-Deixis: deux approches concurrentes” Kleiber Georges. In: Morel Mary-Annick , Danon-Boileau Laurent , eds. La deixis. Colloque en Sorbonne (8-9 juin 1990). Paris: P.U.F.; 1992 .
- “On generative Semantics” Lakoff Georges. In: Steinberg Danny David, Jakobovits Leon A, eds. Semantics. An Interdisciplinary Reader in Philosophy, Linguistics, and Psychology. London & New York: Cambridge University Press; 1971 .
- “From conversational to conventional implicature. The Romanian pronouns of identity and their substitutes” Manoliu-Manea Maria, Carruba Onofrio, Bernini Giuliano. In: Ramat Anna Giacalone, ed. Papers from the 7th International Conference on Historical Linguistics. Amsterdam & Philadelphia: Benjamins; 1987 .
- Discourse and Pragmatic Constraints on Grammatical Choices. A Grammar of Surprises Manoliu-Manea Maria. Amsterdam, Lausanne, New York, Oxford, Shannon & Tokyo: Elsevier; 1994.
- “La pragma-sémantique de l’identité” Manoliu-Manea Maria. In: Mulder Walter de, Ryck Liliane Tasmowski-De, Vetters Carl, eds. Relations anaphoriques et (in)cohérence. Amsterdam & Atlanta: Rodopi; 1997 .
- “Sur l’utilité du mot même” Martin Robert. Travaux de Linguistique et de Littérature.1975;13(2):227-243.
- Pour une logique du sens Martin Robert. Paris: P.U.F.; 1983.
- Etymologisches Wörterbuch der rumänischen Sprache: Lateinisches Element mit Berücksichtigung aller romanischen Sprachen Pușcariu Sextil. Heidelberg: C. Winter; 1975.