Demonstratives in medieval Romania: a comparative analysis in a functional perspective

ABSTRACT

This paper presents a comparative study of demonstratives in the medieval translation in different languages (Latin, Italian, French, Catalan, Spanish and Portuguese) of the same work (an ascetic treatise of Isaac of Nineveh) in a functional perspective. It was confirmed the hypothesis that the demonstrative systems were restructured in the language change process from Latin to Romance languages, not only in terms of forms but especially in terms of functions: the demonstratives began to perform functions which, in the Latin, were expressed by conjunction, explanatory phrase, present participle, relative pronoun, anaphoric and identity pronoun, intensive pronoun and even by the lack of formal resource. Finally, it was found that the context in which there was a major retention of the use of demonstratives is in the expression of immediate contrast.

Historical linguistics; Romance linguistics; Functionalism; Demonstratives; Latin; Romance languages

RESUMO

Neste trabalho, apresentamos um estudo comparado de demonstrativos na tradução medieval em diferentes línguas (latim, italiano, francês, catalão, espanhol e português) de uma mesma obra (tratado ascético de Isaac de Nínive) em uma perspectiva funcional. Confirmou-se a hipótese de que os sistemas de demonstrativos se reestruturaram, no processo de mudanças linguísticas do latim às línguas românicas, não apenas em termos de formas, mas, sobretudo, em termos de funções: os demonstrativos passaram a exercer funções que, no latim, eram expressas por conjunção, locução explicativa, particípio presente, pronome relativo, pronome anafórico e de identidade, pronome intensivo e mesmo por ausência de recurso formal. Por fim, verificou-se que o contexto em que houve maior retenção do uso de demonstrativos foi o de expressão de contraste imediato.

Linguística histórica; Linguística românica; Funcionalismo; Demonstrativos; Latim; Línguas românicas

Introduction

The complexity of the demonstratives already has drawn the attention of linguists for some time (for example, the classic works of Brugmann (1904) and Bühler (1934)) and studies that focus only this linguistic category have recently appeared, but covering data from several languages in order to identify their universal and particular aspects (DIESSEL, 1999; DIXON, 2003).

According to Diessel (1999), the demonstratives (a) are deictic expressions serving to specific syntactic functions (functioning as independent pronouns, noun modifiers or locative adverbs); (b) generally serve to specific pragmatic functions (primarily used to focus the hearer’s attention on an object or on a location in a speech situation, but they may also function to organize the information flow in the ongoing discourse); and (c) are characterized by specific semantic features, contrasting the proximal (reference to the entity close to the deictic center) and the distal (indication that the referent is at some distance from the deictic center), although there are languages in which the demonstrative is neutral in relation to distance.

Although the Romanic nominal demonstrative systems (independent pronouns and noun modifiers) derive from the same matrix (the Latin system), now they show great diversity in their configuration (LAUSBERG, 1981).

We regret that the absence of systematic descriptions, made exactly under the same parameters for Latin and the many Romance languages in its different historical phases complicate significantly the task of even recognizing the effective patterns of organization of the demonstrative systems. (CAMBRAIA; BIANCHET, 2008).

Functionalism

As Neves has explained (1997), functional grammar stands for a theory of grammatical organization of natural languages that fits into an overall theory of social interaction and that considers that grammar is subject to pressures of usage. Neves (1997) has pointed out that the functional paradigm is characterized by the following features: (a) defines the language as a social interaction tool; (b) considers that the primary function of language is communication; (c) has as the psychological correlate the communicated competence, understood as the ability to interact socially through language; (d) states that the linguistic system should be studied within the context of use; (e) requires that the linguistic description provides data to account for their function in a given context; (f) believes that language acquisition is done with the help of an extensive and structured input of data presented in the natural context; (g) explains the language universals based on communication, biological / psychological and contextual constraints; and (h) prioritizes pragmatics, framework within which semantics and syntax are studied.

Moreover, as stated by Martelotta and Areas (2003), premises that are part of the functionalist conception of language in Givón’s view (1995) are: (a) language is a socio-cultural activity; (b) structure serves cognitive and communicative functions; (c) the structure is non-arbitrary, motivated and iconic; (d) change and variation are always present; (e) meaning is contextually dependent and non-atomic; (f) categories are not discrete; (g) the structure is malleable and non-rigid; (h) grammars are emergent; and (i) rules of grammar allow some exceptions.

In addition, it is assumed that linguistic structure derives from discourse and is shaped by it, as in the approach advocated by Votre and Naro:

The fundamental hypothesis of this proposal is that the use of language — communication in the social situation — originates the form of the language, with characteristics that are peculiar, including different degrees of instability associated to different subsystems. This implies to understand language as a malleable, probabilistic and non-deterministic object. Therefore, in this view, structure (or the form of the language) is a dependent variable, resulting from the regularities of situations in which one speaks.¹ (VOTRE; NARO, 1996, p. 51-52, our translation).

A functionalist model that has proven to be especially productive for the study of language change is the typological-functional model of Givón (2001), due to its ability to integrate a functionalist orientation, which emphasizes the communicative function of language in the analysis and is based in the study of language in its context of use, to a typological orientation, which seeks to account for linguistic diversity.

Givón (2001), in the latest version of his theoretical model, classifies demonstratives as forms that fit both in the class of determinants (often unstressed and clitics) and in the class of the independent pronouns (usually stressed and independent)². Givón (2001) has pointed out that certain types of grammatical morphemes may have a sharper differentiation between determinant and independent pronoun (as in French ce × celui). Pronouns — and, therefore, also the demonstratives — are located by Givón (2001) in the intersection of two functional domains: semantic and discourse-pragmatic. From the semantic point of view, pronouns, in its classic paradigm, are grammatical morphemes encoding classificatory traits, among which most commonly person or speech act participants [SAP] (speaker = 1^st person, 2^nd person = listener; non-SAP = 3^rd person), spatial deixis related to SAPs (proximity and/or visibility of the speaker or of the listener), number (singular, dual, plural), class or gender (masculine, feminine, neutral, etc.), and case (subject, direct object, etc.; ergative or absolutive; agent, patient, etc.). Givón (2001) points out, however, that spatial orientation of the demonstratives can be expanded to temporal orientation related to some reference point in time. From the discourse-pragmatic point of view, pronouns are part of the resources that are in the core of the grammar of referential coherence, namely: zero-anaphora, unstressed anaphoric pronouns (in which the demonstratives are included as determinants), stressed independent pronouns (in which the demonstratives are included as pronouns) and definite full NPs.

It is precisely because demonstratives play distinct roles (semantic coding of person and/or space and discourse-pragmatic coding referential coherence) that its system is of so great complexity. One could say that it is a system in permanent “tension” because of the competition between different functional pressures:

That fact that clausal grammar codes simultaneously propositional-semantic information and discourse-pragmatic function has far reaching consequences. For the coding requirements of the two are often in conflict, so that the resulting structure is an adaptive compromise between the competing functional pressures. (GIVÓN, 2001, v.1, p.19, italics of the author).

Although we adopt in this research the typological-functional model of Givón (2001), it should be noted here that we observe the consideration of working with a moderate approach of functionalism, as defended by Votre and Naro:

We do not deny the existence of structure, on the contrary we want to understand their basic motivations, even if these can be exclusively diachronic in certain situations. However, we take this opportunity to note that some functionalists do not share our moderate position, considering [those functionalists] that the structure has no independent existence of language use [...].³ (VOTRE; NARO, 1996, p.52, italics added, our translation).

Working hypothesis

In view of the principles postulated within the framework of functionalism, we will take as a working hypothesis the idea that the demonstrative systems have restructured themselves in the process of linguistic change from Latin to the Romance languages, not only in terms of forms, but especially in terms of functions. Several studies have described these changes from a formal point of view, but comparative studies with empirical basis analyzing the issue from a functional point of view are rare.

Methodology

The comparative study of Romance languages of earlier stages always comes up against many difficulties. Certainly the most important difficulty, considering the functionalist perspective, is the identification of texts of the same nature or at least of strongly similar nature: how can one make a precise comparison from the functional point of view if, in different texts, contexts where the functions are present are also different? Due to that, an interesting approach is to work with parallel texts, that is, texts with the same content but in different languages, case in which the translations are the most typical examples.

This study has followed this orientation: the corpus is constituted by medieval translations in Latin, Italian, French, Catalan, Spanish and Portuguese of the so called Book of Isaac. This work consists of an ascetic treatise written by the anchorite Isaac of Nineveh (7^th century) in Syriac, which was then translated into Greek (around 8^th century), from this translation into Latin (by the 3^rd quarter of 13^th century), and then into the Romance languages (CAMBRAIA, 2000). Since there is no critical edition of this work for all mentioned languages, we have chosen here to analyze the demonstratives in just one testimony of each language (except in the case of Spanish, as it will be explained below). The adopted testimonies were as follows (all are manuscripts, except the one of Seville, which is a printed edition)⁴:

1. Latin (L): Milan, Pinacoteca Accademia Ambrosiana Library, A 49 sup, 13^th cent., ff. 1^r-75^v;

2. Italian (I): Florence, Riccardiana Library, Ricc. 1489, 14^th cent., ff. 10^r-155^v;

3. French (F): Paris, National Library, Lat. 1489, 15^th cent., ff. 308^r-365^v;

4. Spanish-1 (E1): Madrid, Royal Palace Library, II/795, year 1484, ff. 1^r-123^r;

5. Catalan (C): San Lorenzo de El Escorial, Royal Library of the Monastery, n.I.16, 15^th cent., ff. [0^r]-69^r;

6. Spanish-2 (E2): Seville, year 1497, ff. 127^v-162^v;

7. Portuguese (P): Rio de Janeiro, National Library, 50-2-15, 2^nd half of the 15^th cent., ff. 1^r-114^r.

We must present some information about the Latin-Romance tradition of the Book of Isaac in order to enable a better understanding of what these testimonies represent.

The handwritten Latin tradition is composed of approximately 100 manuscripts, and L is one of the oldest and the one which is the most faithful to the original text (CAMBRAIA; LARANJEIRA, 2010). The Italian tradition consists of approximately 25 manuscripts, and I is the most complete and faithful among the oldest (VILAÇA, 2012). The French tradition consists of only one manuscript (MELO, 2010): F. The Catalan tradition comprises 3 testimonies (with two independent translations, one of them fragmentary): C is the most complete testimony (CAMBRAIA; CUNHA, 2008). The Spanish tradition is divided in two independent translations, one – E1 – probably is derived from the Latin translation and the other – E2 – from the Catalan. The Portuguese tradition is composed of 4 manuscripts: P is the most faithful and complete, and is derived from the Spanish translation (the Portuguese tradition is bound to the tradition of E2, but not directly to this testimony). The traditions in Italian, French, Catalan and Spanish (E1) are derived directly from Latin (though not necessarily of the same Latin testimony). See below a simplified representation of the genetic relationship⁵ between these testimonies, adapted from Avellar (2011, p. 4), where *L (= hypothetical Latin testimony/ies), *C (=hypothetical Catalan testimony) and *E (= hypothetical Spanish testimony):

Figure 1
– Simplified stemma

Source: Author’s elaboration embased on Avellar (2011, p.4).

It is important to clarify that this simplified stemma is intended only to show the relationship between the 7 testimonies used as the source for the present study: it is estimated that there are several intermediate testimonies (hypothetical and remaining) among the stated ones, but even though the stemma allows to see clearly their relationship in terms of branches of tradition.

The stemma also shows the need to always consider the interpretation of data depending on the type of influence of the model: in I, F, E1 and C, there is a supposed influence of the Latin model, but in E2 and in P the influences would be respectively of the Catalan model and the Spanish model. It should be also noted also that, given the profusion of Latin testimonies of this work in medieval Europe, one has to consider the Latin influence also in the latter two languages, by contamination in the tradition, that is, simultaneous access to models in Romance language and in Latin (CAMBRAIA, 2005).

Of course these considerations immediately cast doubt on the productivity of working with translated texts, but to this objection can be argued that: (a) the interference of Latin in the Middle Ages was widespread and this is not an isolated case⁶ and (b) the system of demonstratives in these texts, albeit with Latin influence, should represent, even partially, the system in use, otherwise the texts obviously would not be understood by readers of that time.

Given the complexity of working with such particular texts, a sample of the collection was imposed: we have collected all the data through the first chapter of the Latin translation (it consists of approximately 3300 words) and the respective chapters in Romance translations (which appear generally divided into more than one chapter). The excerpts appear transcribed conservatively⁷ (CAMBRAIA, 2005), but with development of abbreviations (without its indication with italics) and with inclusion of word separation with space or apostrophe, both procedures to facilitate the reader’s job.

Again, given the complexity of the data, it was necessary to differentiate contrastive demonstratives [= CDs] (those that oppose themselves in terms of space and/or person, as those of the Portuguese system with este/esse/aquele) and not contrastive demonstratives [= NCDs] (those that do not take part in the aforementioned oppositive system, such as Portuguese o): it was considered NCDs the forms ciò in Italian, ce in French, çò in Catalan, el/lo in Spanish and o in Portuguese. A second fundamental difference between CDs and NCDs is the possibility of expression of exophora and endophora by the former, while the latter express only endophora (either as anaphora, either as cataphora).

Data collection had as reference the CDs and their inflections: Latin hic/iste/ille; Italian questo/codesto/quello; French cist/cil; Catalan aquest/aqueix/aquell; Spanish este/ese/aquel; Portuguese este/esse/aquele⁸ – and their reinforced variants (as French icil, Spanish aqueste, Portuguese aqueste, etc.). For each occurrence of these forms in the corpus, we sought the corresponding expression in the other translations, even if they were not being expressed with CDs – this is, by the way, one of the central tasks of this study: to know the different formal expressions of the same function.

Data description

Applying the method of data collection described above, it was possible to identify 244 places in the corpus where one or more language had a form of CD.

Inventory

In this section we present the inventory of CDs collected in the corpus of this research, adding some necessary comments. In the tables below, the subscript number after a form is the number of occurrences in the corpus. As for L, E1, C, E2 and P there is no different systems for determinant and independent pronoun (except for neutral in E1, C, E2 and P, which is always an independent pronoun), we did not include this distinction in the tables: we inform in note, however, the forms of the category that is less frequent.

Table 1
– Medieval Latin translation9: L

	Determinant10/Independent pronoun
	m.s.	f.s.	n.s.	m.p.	f.p.	n.p.	Total
F1	hic [nm.]3, huius [gn.]5, hoc [ab.]2	hec [nm.]1, hac [ab.]1	hoc [nm.]4, hoc [ac.]2, huic [dt.]1, hoc [ab.]4	−	−	hec [nm.]1, hec [ac.]4, hiis [dt.]1, hiis [ab.]4	33 (59%)
F2	iste [nm.]1	−	istud [nm.]1	−	−	istorum [gn.]1	3 (5%)
F3	ille [nm.]3, illum [ac.]1, illi [dt.]2, illo [ab.]2	illam [ac.]2, illa [ab.]3	illud [nm.]1, illud [ac.]2	illos [ac.]1, illis [dt.]1	−	illa [ac.]1, illorum [gn.]2	20 (36%)

• Source: Author’s elaboration.

In Latin, there was an instance of hic that was not computed in Table 1 because, as a matter of fact, it is an adverb (in the function of contrast between illic/hic), as seen in the excerpt 1 below:

Exc. 1

L: illic enim sollicitudo est necessaria hic uero dilatatio cordis (f. 1v20) [“there then (where) care is needed, here (there is) truly the extention of heart”]¹¹

It is interesting to note in Table 1 the existence of the use of F2 in L only without its traditional value of reference to the listener: we note the result of a trend that had been marked by Keller (1946) in Classical Latin.

Table 2
– Medieval Italian translation: I

	Determinant12/Independent pronoun				Independent pronoun
	m.s.	f.s.	m.p.	f.p.	m.s.	f.s.	m./f.p.	Total
F1	questo 13	questa 4	−	queste 7	questi 4	−	−	28 (35%)
F2	−	−	−	−	−	−	−	−
F3	quel2, quello5	quella 7	quelli 2	quelle 10	quegli4, colui12	−	coloro 11	53 (65%)

• Source: Author’s elaboration.

It is interesting to note for Italian the fact of not appearing any occurrence of the forms of F2, namely codesto and inflexions.

Table 3
– Medieval French translation13: F

		Determinant				Independent pronoun					Total
		m.s.	f.s.	m. p.	f.p.	m.s.	f.s.	n.s.	m.p.	f.p.
FI	NC	cis 1	ceste 1	−	ces 2	cis 2, cis ci 3	ceste 1	−	−	−	10 (14.3%)	22 (31.4%)
	DOC	−	−	−	ces 3	−	−	est 1	−	−	4 (5.7%)
	IOC	cest 2	ceste 2	−	ces3, ices1	−	−	−	−	−	8 (11.4%)
FII	NC	−	−	−	−	cil4,celi(?)1,celui1	−	−	−	ical(?)1	7 (10%)	48 (68.6%)
	DOC	−	cele2, celle2	−	−	celui 1	−	−	ceuls 1	celles1, icelles1	8 (11.4%)
	IOC	cel 2	celle 6	−	−	celi(?)2,celui3, icelui1	celle 1, icelle 3	−	cil1, ceuls7,iceuls2,iceus1	celles1, icelles3	33 (47.1%)

• Source: Author’s elaboration.

With regard to French, there was some data, included in the respective tables, which are uncommon. There are 3 instances of celi as masculine singular independent pronoun, but the form celi in medieval French was used for feminine singular in the indirect regime case (MARCHELLO-NIZIA; PICOCHE, 1998): it is possible that it is a mistake of the scribe, who omitted the u of the masculine form celui, as seen in the excerpt below:

Exc. 2

F: Aussi comme cil qui ne uoit le soleil de ses yex ne puet a aucun fors de son oie seulement reciter la lumiere de celi (f. 312v20-22) [“As one who does not see the sun with his eyes can to no other only by his ear describe the light of that (sun)”]

Since this form occurred 3 times, there may be another explanation: in the disappearance process of celi (absent of modern French), it is possible that this has been confused with celui, being that interpreted as a phonological variant of this.

Another unforeseen occurrence is the one of cil as masculine plural as independent pronoun in the indirect regime case, although it usually serves to subject case, the most common being ceuls to (direct or indirect) regime case:

Exc. 3

F: Ce est a savoir des fausaires et de cil qui vendent les devins parlemens (f. 311v10-11) [“Namely: of the counterfeiters and of those who sell the divine words”]

Also for this case, it seems to be possible the two previous explanations: mistake of the scribe or syncretism between forms (since cil also disappeared).

The occurrence of ical for feminine plural in the subject case is also unusual, because the expected form would be celles or the reinforced icelles: it may have been a confusion of the scribe.

Exc. 4

F: Mais se mistrent en la mer de cest siecle a sauuer les ames des autres comme ical fussent encore malades et perdirent eus meismes de l’esperance de dieu (f. 310v19-22) [“But (they) were cast into the sea of this world to save the souls of the others, since those were still ill, and have lost themselves of the hope of God”]

It is also worth noting the fact that there is already record of the combination of demonstrative and adverb, as in the case of cis ci, a feature that would grammaticalize in the course of the history of French language. It is interesting to note that, as evidenced by Dees (1971) in relation to his corpus, the first pattern of combination is between forms of same content ‒‒ either proximity (as in cis ci), either distance ‒‒ up to approximately 1350. It is a curious fact because Ouy (1999, t. 1, p. 303) has proposed approximately 1425 as date for F: we assumed then that the language of F is conservative, as it is common in religious texts.

Finally, there is the form est, which is curious, since French forms usually present a c- as a remain of the Latin ecce of reinforcement. It is probably an archaism, since it appears in very ancient texts such as La Vie de Saint Alexis (11^th cent.): “De tot est mond somes nos jugedor” (GODEFROY, 1885, v.4, p. 618, italics added).

Table 4
– Medieval Spanish translation: E1

	Determinant3 (except n.s.)/Independent pronoun
	m.s.	f.s.	n.s.	m.p.	f.p.	Total
F1	este 11	esta 7	esto21, aquesto1	−	estas 5	45 (48%)
F2	−	essa 2	esso 1	essos 1	−	4 (4%)
F3	aquel 17	aquella 7	−	aquellos 8	aquellas 13	45 (48%)

• Source: Author’s elaboration.

It is worth noting the presence of an occurrence of reinforced neutral form aquesto: perhaps this is the context of greater resistance for its loss.

Table 5
– Medieval Catalan translation: C

	Determinant15 (except n.s.)/Independent pronoun					Total
	m.s.	f.s.	n.s.	m.p.	f.p.
F1	aquest 12	esta1, aquesta5	aço 11	−	aquestes 13	42 (40%)
F2	−	−	−	−	−	−
F3	aquel2, aquell16, cell7	aquella 6	allo 3	aquells11, cells6	aquelles 12	63 (60%)

• Source: Author’s elaboration. ¹⁹

There are three aspects that stand out in the Catalan data: the absence of F2 (aqueix and inflexions), the presence of the not reinforced form (esta) and the presence of the forms cell and cells. The prevalence of reinforced forms (aquest, etc.) suggests that the language is from the final archaic phase, since in the beginning of it the reinforced forms were less frequent. The absence of the forms of F2 seems to be related to the question of textual genre, as we will comment below, in the end of this section. Finally, the forms cell and cells, in 13 occurrences, have the specificity of always appearing as introducer of relative clause (for example: “cell qui uol nobles uestirs no pot auer humjls cogitacions”). Although Badía i Margarit (1994, p.313) mentions the existence of oppositive related form (cest), this one did not appear in the corpus.

Table 6
– Medieval Spanish translation: E2

	Determinant16 (except n.s.)/Independent pronoun
	m.s.	f.s.	n.s.	m.p.	f.p.	Total
F1	este6, aqueste2	esta 4	esto51, aquesto1	−	estas 9	73 (54%)
F2	−	−	eso2, esso4	−	−	6 (4%)
F3	aquel 27	aquella 5	aquello 1	aquellos 17	aquellas 6, aquallas 1	57 (42%)

• Source: Author’s elaboration. ²⁰

Table 7
– Medieval Portuguese translation: P

	Determinant17 (except n.s.)/Independent pronoun
	m.s.	f.s.	n.s.	m.p.	f.p.	Total
F1	este 11	esta 6	sto5, esto37, aquesto1	−	estas 9	69 (57%)
F2	−	−	eso1, esso3	−	−	4 (3%)
F3	aquel18, aquell3, aquele1, aquelle2	aquela 4, aquella 1	aquelo 1	aquelles 17	aquellas 3	48 (40%)

• Source: Author’s elaboration. ²¹

Data from the second Spanish translation (Table 6) and from the Portuguese one are very similar, with remains of reinforced forms and forms of F2, as in other Spanish translation. In the case of reinforced forms, P shows the same pattern of E1 with only neutral forms, while E2, despite having one form of neutral, presents two forms of masculine.

It is interesting to note that in all instances of F2 in P and in almost all in E2, the demonstrative appears accompanied by mesmo/meesmo, suggesting that they are semantically related to the idea of intensiveness.

Comparing the tables, it is possible to note firstly that, in general, F2 has very little expression in L, E1, E2 and P and does not occur in I and C (never existed in F). One possible explanation for the low or zero frequency is the rare reference to listener in the text in this kind of textual genre: being an ascetic doctrinal text, there is virtually no opportunity for reference to the listener nor for dialogues.

A second interesting aspect is the difference in preference between the forms of proximity and/or 1^st person (F1 and FI) and the ones of distance and/or 3^rd person (F3 and FII): I, F and C make more frequent use of the former; and E2 and P, of the latter. For E1, there is no preference¹⁸. A possible generalization would be that the binary demonstratives systems would favor the forms of distance and/or 3^rd person, while the ternary one would favor the forms of proximity and/or 1^st person or, put in a more functional perspective, the most common would be the unmarked forms of each system.

Patterns of relation

In the data collected in the corpus there are basically 3 large patterns of relation.

(a) Pairing of the same class: the presence of the same excerpt with CD in all translations.

Exc. 5

L: Labores huius seculi qui pro ueritate fiunt non comparantur delicijs (f. 1r14-15)

I: Gli exercitii di questo secolo li quali si fanno per uanitade. Non si aguaglano alle delitie (f. 10v13-15)

F: Li labeurs de cest siecle qui sont fait pour la nesescite du cors ne soient mie compare aus delices (f. 308v33-35)

E1: Los trabaios que en este siglo por la verdat se çufren; no tienen comparacion con los gozos y deleytes (f. 2r21-2v1)

C: Los trebaylls dequest secgla qui son fets per uerjtat no son comparables als deljts (f. 0vb21-22)

E2: Los trabajos deste mundo no son comparados a los deleytes (f. cxxvij-vb4-5)

P: E os trabalhos deste mundo nõ som conparados aos deleitos (f. 4r16-17)

(b) Pairing of different classes: the presence of the excerpt with CD in one or more translations, but not in all, with expression of the content through other explicit formal resource.

Exc. 6

L: Sicut secuntur seminantes in lacrimis manipuli exultationis (f. 1r16-17)

I: sicome ad coloro che seminano in lagrime seguitano bracciate di gioconditade (f. 10v16-18)

F: Auci comme les manieres des leecemens ensieuent les semans en lermes (f. 308v36-38)

E1: Ca assi como aquellos que siembran en lagrimas e lloros coxen fasces de soberano gozo (f. 2v3-6)

C: axi com cells quj sembren en lagrames conseguexen guardo de gran alegria (f. 0v25-27)

E2: assi como aquellos que siembran en lagrimas alcançan galardon de grande alegria (f. cxxvij-vb7-9)

P: assy como aquelles que semeã as lagrimas alçãçam galardom de grande alegria (f. 4r18-19)

Note that, in L and F, there is no demonstrative as in the other translations, rather a present participle (lat. seminantes and fr. semans).

(c) Pairing with gaps: the presence of the excerpt with CD in one or more translations, but not in all, with cases without expression of the content through other explicit formal resource, as in the following case¹⁹:

Exc. 7

L: et comprehendit illam rem pro qua christus aduenit Ø (f. 8v4-5)

I: et ae conpreso quella cosa per la quale uenne christo Ø (f. 24r1-2)

F: et comprent celle chose pour la quelle ihesu crist vint Ø (f. 314v8-9)

E1: e comprehende aquella cosa por la qual hiesu christo vino en el mundo (f. 14r-12-13)

C: E a aquella cosa consegujda per la qual nostre senyor Jhesu christ uench en aquest secgla (f. 6vb15-17)

E2: e ha fallado aquella cosa: por la qual nuestro señor ihesu christo vino en este mundo (f. cxxxi-ra40-42)

P: e ha achada aquela cousa pola qual o nosso Senhor Jhesu christo ueeo ẽ este mundo (f. 14v4-5)

Note that, in L, I and F, there is not the excerpt in which appears the demonstrative in C (aquest), E2 (este) and P (este). In E1, on the other hand, there is the excerpt, but in the place of DC there is the article el.

Pairing of the same class

Of the 244 points in the text with CD in one or more translations, only in 21 (9%) there is pairing of the same class. These data are distributed into 3 subtypes: 10 occurrences of pairing with the proximal form (F1 and FI), as in Exc. 5 above; 10 occurrences of pairing with the distal form (F3 and FII) as in Exc. 8 below; and 1 occurrence of pairing with variation (distal in I but proximal in the others), as in Exc. 9 below.

Exc. 8

L: et consolationem illam non sentiet de qua erat apostolus consolatus (f. 8r19-20)

I: Ne non sentira quella consolatione de la quale era consolato L’apostolo (f. 23v4-5)

F: et ne sentira cele consolation de la quel li apostre estoit consolles (f. 314r36-37)

E1: ny sentira aquella consolacion de la qual era ell apostol sant pablo aconsolado (f. 13v16-17)

C: ne sintra aquella consolacio daquell apostol sent paul era consolat (f. 6va18-19)

E2: ni sentira aquella consolacion de la qual era consolado el Apostol (f. cxxxi-ra17-19)

P: nẽ sentira aquela cõsolaco da qual era consolado o apostolo (f. 14r10-11)

Exc. 9

L: Attende igitur o homo hec que legis (f. 8v2-3)

I: Adumque o homo considera quelle cose che tu leggi (f. 22v19-20)

F: Entent donques os tu homs ces choses que tu lis (f. 314r15-16)

E1: Para mientes pues o hombre en esto que lees (f. 13r15-16)

C: O hom guarde ben e entin aquestes coses que ligs (f. 6rb23-25)

E2: O hombre guarda e entiende bien estas cosas que lees (f. cxxx-vb33-34)

P: Oo homẽ esguarda e entende bem estas cousas que lees (f. 13v9-10)

The essential question is: why exactly in these 21 cases was there the maintenance of CD in all translations? Or, rather, which functions are these that require the specific use of CD for its expression?

The 21 occurrences are distributed between 3 functions: (a) spatial deixis expressing proximity to the speaker (see Exc. 5), with 2 occurrences ; (b) cataphora (relative clause introducer) (see Excs. 8 and 9), with 10 occurrences ; and (c) anaphora, with 8 occurrences . It is interesting to note that, of 18 occurrences of the latter two categories, 7 (4 cataphora and 3 anaphora) are in the context of immediate contrast, that is, there is an opposition with different resources between sentences that follow one immediately after the other (be in two consecutive sentences, be in two consecutive clauses of the same sentence). See in Chart 1 below (we have marked with superscript R the terms that are introducers of relatives):

Chart 1
– Expression of the function of immediate contrast20

	Exc. 10	Exc. 11	Exc. 12	Exc. 13	Exc. 14	Exc. 15
L	hec... ab illa	hic... ille R	isR... illeR	iste... illo R	qui... illi R	facientes... illis R
I	questa... da quella	questi... colui R	coluiR... coluiR	questo... colui R	chi... quelli R	coloroR... coloroR
F	ceste... d’icelle	cis ci... celui R	cilR... celiR	cis ci... celi R	qui... ceuls R	faisans... ceuls R
E1	esta... de aquela	este tal... aquel R	el ciego R ... aquel R	este... aquel R	quien... aquellos R	aquellosR... aquellosR
C	aquesta... dequella	aquest... cell R	aquellR... dequellR	aquest... dequell R	cell R ... aquells R	aquellsR... aquellsR
E2	esta... da aquella	este... Ø	aquelR... de aquelR	Ø... de aquelR	Ø... aquellosR	aquellosR... aquellosR
P	esta... daquela	este... Ø	aquellR... daquellR	Ø... daquelR	Ø... aquellesR	aquellesR... aquellesR

• Source: Author’s elaboration.

Exc. 15 (chart 1)

L: Noli comparare facientes signa et prodigia et virtutes in mundo illis qui in solitudine sunt scienter (f. 3r8-10)

I: NOn aguaglare coloro che fanno isegni et lemarauegle et le uirtudi nel secolo ad coloro che sono sauiamente insolitu dine (f. 14v3-7)

F: Ne ueilles mie comparrer les faisans signes merueilleus et vertus en monde a ceuls qui sont sagement en solitude (f. 310r29-31)

E1: No fagas comparacion de aquellos que fazen senyales mílagros e virtudes en el mundo con aquellos que discretamente y como deuen moran en la solitut (f. 5r17-19)

C: No vulles comparar aquells qui en lo mon fan grans senyals e mjracles a aquells qui sientalment son en soljtut (f. 2ra30-2rb2)

E2: No quieras ygualar a aquellos que en el mundo fazen milagros e virtudes e grandes marauillas con aquellos que estudiosamente estan en apartado (f. cxxviij-va21-22)

P: Nom queiras Jgualar aquelles que en no mundo fazẽ milagres e uirtudes e grandes marauilhas cõ aquelles que estudosamẽte estam ẽ apartado (f. 7r2-4)

Of the data presented in chart 1, the 7 occurrences of pairing of the same class are: the two terms of Exc. 10, the first term of Exc. 11 and the second term of Excs. 12 to 15.

Since, even in this context of immediate contrast, there is no categorical use of CDs (check the first term of the Exc. 6), one must necessarily understand its use as a variable phenomenon. It is possible to imagine that, since in the context of contrast it can also appear the need of the relative clause introducer, there would be then an overlap of motivations leading to a major preference for the use of CD in this case: this overlap is present in the paired terms of Excs. 11 to 16.

Pairing of different classes

Of the 244 points of the text with CD in one or more translations, 115 (47%) presented pairing of different classes.

A first point of interest is to evaluate the productivity of CDs for expression of the set of functions comprehending these 115 points of the text. The data are distributed as follows:

Table 8
– Resources for the expression of the same set of different functions

	L	I	F	E1	C	E2	P
CD	23 (20%)	45 (39%)	28 (24%)	39 (34%)	52 (45%)	82 (71%)	73 (63%)
NCD	−	24 (21%)	27 (23%)	13 (11%)	30 (26%)	3 (3%)	4 (3%)
Other resources	92 (80%)	46 (40%)	60 (52%)	63 (55%)	33 (29%)	30 (26%)	39 (34%)

• Source: Author’s elaboration.

The first interesting result of the comparison is that demonstratives in general (CD and NCD) are more frequent in Romance languages than in Latin.

A first explanation for this Romance profusion of demonstratives is partially in the history of Latin conjunction (Cj): most conjunctions were monosyllabic and disappeared, being their function of inter-clausal articulator assumed by conjunctive phrases containing demonstratives.

Table 9
– Correspondence in the function of inter-clausal articulator (coordinate and/or adverbial clauses)

	L	I	F	E1	C	E2	P
Cj	27 (100%)	−	−	−	−	−	−
CD	−	−	−	−	−	25 (93%)	19 (70%)
NCD	−	21 (78%)	4 (15%)	−	22 (81%)	−	−
Other resources	−	6 (22%)	23 (85%)	27 (100%)	5 (19%)	2 (7%)	8 (30%)

• Source: Author’s elaboration.

Exc. 16

L: Perseuera legens in solitudine ut mens tua semper ad dei mirabilia de[d]ucatur (f. 2r24-25)

I: Perseuera in solitudine leggendo adcio ke la tua mente sempre sia menata alle marauiglose cose di dio (f. 13r5-8)

F: Perceuere en solitude lisant si que ta pensee soit toudis demenee aus merueilles de dieu (f. 312r8-9)

E1: Lee de contino en la soledat porque sea tu piensa siempre occupada en las marauillas de dios (f. 4r11-12)

C: perseuera ligent en la solitut per ço que la tua pensa sie tostemps endressade en les mereuelles de deu (f. 1va28-30)

E2: Perseuera leyendo en apartado por esto que el tu pensamiento sea todos tiempos traydo en las marauillas de dios (f. cxxiij-rb8-11)

P: Perseruera senpre leendo em apartado por tal que teu penssamento seia todos tenpos tragido en nas marauilhas de deus (f. 6r1-3)

Table 9 refers to the conjunctions and conjunctive phrases that act as inter-clausal articulators for coordinate and/or adverbial clauses in the Latin data of pairing in cases of different classes and the corresponding forms in other translations. It is remarkable how the demonstratives were employed to play the function of articulator in Romance languages: in the data of L there isn´t any occurrence of demonstrative in the function of articulator, while in Romance languages there is a clear prevalence in I, C, E2 and P, although not in F and E1. Of 27 occurrences of articulators in L, no less than 16 are the final conjunction ut (that did not survived in any of the Romance languages): to express this function, the preferred way in each translation is adcio/accio que/ke in I, si que in F, porque in E1, per ço que in C, por esto que in E2 and P. A significant difference among the data of Romance languages is the fact that NCD is the preferred demonstrative for the function of articulator in I, F and C, while in E2 and P it is CD: since the function of articulator is expressed by the conjunctive phrases with the preposition, one can understand that NCD is not preferably employed in E2 and P, since it is unstressed monosyllabic and, therefore, incompatible with the function of prepositional complement.

A second explanation is in the expression of the function of introducer for clarification, which in most Romance languages was restructured with a demonstrative. To show this, we present a new table, now differentiating the data with these functions in Latin and their corresponding in Romance languages.

Table 10
– Correspondence in the function of introducer for clarifying

	L	I	F	E1	C	E2	P
CD	−	−	−	1 (50%)	−	2 (100%)	2 (100%)
NCD	−	2 (100%)	2 (100%)	−	2 (100%)	−	−
Other resources	2 (100%)	−	−	1 (50%)	−	−	−

• Source: Author’s elaboration.

Exc. 17

L: nisi fiat misericors supra iusticiam non est misericors id est quod hominibus non solum misereatur de propriis (f. 2r4-6)

I: se lhomo misericordioso non e piu che giusto non. e misericordioso cioe che non solamente faccia misericordia altrui dele proprie cose (f. 12r12-16)

F: se tu nes mesericors outre iustice. Cest que nemie seulement tu faces misericorde de ton propre (f. 312r8-9)

E1: [si el miseri]cordioso no anduuiere sobre la iusticia fun[dado no] es misericordioso. Esto es dezir que no solo faga misericordia en los hombres de lo proprio (f. 3v5-7)

C: si no es misericordios sobre Justicia ell no es misericordios ço es que no tant solament aye merce de coses propries als altres (f. 1rb26-29)

E2: si no es misericordioso sobre la iusticia que no es misericordioso. Esto es que no tan solamente fagas merced e limosna de las cosas propias a los otros (f. cxxiij-ra20-24)

P: sse nõ he mjsericordoso sobre a Justiça que nõ he misericordoso esto he que nom tam solamente aya mercee de cousas proprias aos outros (f. 5r22-5v1)

In data of L there is no instance of demonstrative as introducer for clarification while in Romance languages there is a clear prevalence in I, F, C, E2 and P, although not in E1. Again one can see the fact of NCD being the preferred demonstrative for introducer for clarification in I, F and C, while for E2 and P is CD: unstressed NCD would be avoided, since it would take the function of nucleus of the subject of the verb to be.

A third fact that would explain the lower frequency of demonstratives in Latin in relation to Romance languages is related to the expression of the function of agentive, presented in L mainly through present participle (PrP). Interestingly, although PrP have survived in Romance languages as a form that is more nominal than verbal, it seems there is no recruitment of it as frequently as in Latin. In several data of the corpus there is an opposition between the PrP in Latin and a structure with demonstrative (mainly followed by relative) in one or more Romance languages.

Table 11
– Correspondence in the function of agentive

	L	I	F	E1	C	E2	P
PrP	12 (100%)	−	7 (58%)	−	1 (8%)	1 (8%)	−
CD	−	10 (83%)	3 (25%)	7 (58%)	11 (92%)	11 (92%)	11 (92%)
NCD	−	−	2 (17%)	3 (25%)	−	−	1 (8%)
Other resources	−	2 (17%)	−	2 (17%)	−	−	−

• Source: Author’s elaboration.

In Table 11, it is seen that 12 occurrences of L with PrP correspond preferably to structures with demonstrative (followed by relative), except in F. See below a group of data illustrating the different types of correspondence:

Exc. 18

L: ne nolentibus²¹intueri res breviores (f. 5r26-27) [PrP]

I: adcio che coloro che uoglion uedere le cose suttili (f. 18r20-21) [CD + Relative]

F: pour ce que eus veuilians regarder le plus brief chose (f. 312r8-9) [PrP]

E1: porque a los que las mas breues y menores cosas quisieren profundamente veer (f. 9r5-6) [DNC + Relative]

C: per ço que cells qui encare no saben les coses menors ne pus baxes (f. 4ra23-25) [CD + Relative]

E2: Por esto que aquellos que no saben: ni han comprehendido las menores cosas (f. cxxix-va38-40) [CD + Relative]

P: E por que aquelles que nõ sabem nẽ conprendem as meores cousas (f. 10r7-8) [CD + Relativa]

A fourth fact is related to the expression of the function of inter-clausal articulator for substantive clauses, expressed by the relative pronoun (RP) in free relatives (without antecedent) in L, often corresponding to demonstrative with relative in Romance translations.

Table 12
– Correspondence in the function of inter-clausal articulator (substantive clauses)

	L	I	F	E1	C	E2	P
RP	13 (100%)	5 (48%)	6 (46%)	5 (48%)	2 (15%)	2 (15%)	2 (15%)
CD	−	8 (62%)	3 (23%)	1 (8%)	8 (62%)	10 (77%)	10 (77%)
NCD	−	−	3 (23%)	6 (46%)	3 (23%)	1 (8%)	1 (8%)
Other resources	−		1 (8%)	1 (8%)	−	−	−

• Source: Author’s elaboration.

In Table 12, it is seen that the 13 occurrences of L with RP correspond preferably to structures with demonstrative (followed by free relatives), again except in F. See below a group of data illustrating the different types of correspondence:

Exc. 19

L: qui splendida diligit humiles cogitationes habere non potest (f. 2v15-16) [RP]

I: ki ama li splendidi uestimenti non puote auere humili cogitationi (f. 13v13-15) [RP]

F: qui aime choses resplandisans il ne puet auoir humbles cogitacions (f. 310r6-7) [RP]

E1: el que las cosas ricas e fermosas dessea y codicia: no puede hauer humildes cogitaciones (f. 4v12-13) [NCD + Relative]

C: cell qui uol nobles uestirs no pot auer humjls cogitacions (f. 1vb26-28) [CD + Relative]

E2: aquel que quiere aver nobles vestiduras no puede auer humildes cogitaciones (f. cxxiij-rb35-37) [CD + Relative]

P: aquel que quer auer nobres vestiduras nõ pode auer homjldosas cuydacones (f. 6r23-6v1) [CD + Relative]

A fifth fact which would explain the lower frequency of demonstratives in Latin in relation to Romance languages is related to the expression of the function of anaphora made by the anaphoric pronoun (AP) is/ea/id and their identity pronouns (IdP) correlates idem/eadem/idem (FARIA, 1958): in 11 cases, the presence of this form in Latin corresponds often to demonstratives in Romance languages.

Table 13
– Correspondence in the function of anaphora

	L	I	F	E1	C	E2	P
PA/PId	11 (100%)	3 (27%)	3 (27%)	5 (45%)	2 (18%)	1 (9%)	2 (18%)
CD	−	3 (27%)	5 (45%)	−	4 (36%)	5 (45%)	3 (27%)
NCD	−	−	2 (18%)	5 (45%)	2 (18%)	1 (9%)	1 (9%)
Other resources	−	5 (46%)	1 (9%)	1 (9%)	3 (27%)	4 (36%)	4 (36%)

• Source: Author’s elaboration.

In Table 13, it is seen that 11 occurrences of L with AP/ IdP correspond preferably to structures with demonstratives (not necessarily followed by relative clauses) in F, C, E2 and P, except in I and E1. See below a group of data illustrating different types of correspondence:

Exc. 20

L: et ab eis elongauerit se omnino (f. 8r24-25) [AP]

I: et da esse al postutto si dilunghera²² (f. 23v13) [PP3]

F: et sera de eus du tout eslongie (f. 314r1) [PP3]

E1: e del todo dalli adelante se apartare dellos (f. 14r2-3) [PP3]

C: es sera ben lunyat dequells de tot (f. 6va29-30) [CD]

E2: e apartara assi mesmo de aquallas de todo en todo (f. cxxxi-ra28-29) [CD]

P: e apartar sy meesmo dellas de todo en todo (f. 14r19) [PP3]

Finally, the sixth relevant fact which would explain the lower frequency of demonstratives in Latin in relation to Romance languages is the intensive pronoun (IP) ipse/ipsa/ipsum (FARIA, 1958): in 9 occurrences, the presence of this form in Latin corresponds often to demonstratives in Romance languages:

Table 14
– Correspondence in the function of intensiveness

	L	I	F	E1	C	E2	P
IP	9 (100%)	−	−	−	−	−	−
CD+IP	−	−	−	2 (22%)	−	−	−
CD	−	2 (22%)	7 (78%)	3 (33%)	4 (44%)	4 (44%)	2 (22%)
NCD	−	−	1 (11%)	−	−	−	−
Other resources	−	7 (88%)	1 (11%)	4 (44%)	5 (56%)	5 (56%)	7 (78%)

• Source: Author’s elaboration.

In Table 14, it is seen that the 9 occurrences of L with IP correspond preferably to structures with demonstratives in F and E1, but not in I, C, E2 and P. See below a group of data illustrating the different types of correspondence:

Exc. 21

L: et postmodum ipsi qui uiuificarunt alios ad execrabilia vicia et horribilia ceciderunt (f. 3r28- 3v2) [IP]

I: et poi essi che anno uiuificati gl’altri sono caduti in cose maligne et orribili (f. 15r7-9) [Other resource:PP3]

F: et apres ce qu’il ont viuifies les autres sont cheus en pechies aueulables et orribles (f. 310v14-15) [NCD]

E1: Empero despues al fín essos mismos que dieron vida a los otros cayeron en detestables vicios spantables e abhomínables peccados (f. 5v20-6r1) [CD+IP]

C: E puys ells qui aujen los altres endressats cahegueren en greus ujcis e an greus peccats (f. 2va4-6) [Other resource: PP3]

E2: e despues ellos mesmos en feos pecados cayeron (f. cxxviij-vb7-8) [Outro recurso: Other resource: PP3+IP]

P: depois elles meesmos ẽ feos peccados cairam (f. 7v2) [Other resource: PP3+IP]

Pairing with gap

Of the 244 points in the text with demonstratives in one or more translations, 108 (44%) present pairing with gap.

The gaps allow considering the hypothesis of cases of the pairing of same class disrupted by the absence of data in one or more translations; however, among the 11 cases of CD in L, only in one such hypothesis would be valid, since in the others, although there is gap in one or more translation, another resource is used, excluding these data from a hypothetical pairing of same class.

Therefore, cases of pairing with gap are actually cases of pairing of different class, with the particularity of having gap in one or more translations. The patterns repeat in a way what was verified in the previous section: there is a greater use of demonstratives (CD and NCD) in Romance languages as a correspondence, in L, to inter-clausal articulator of coordinates and/or adverbial clauses (7 cases), introducer of clarification (5), agentive (3), inter-clausal articulator (substantive clauses) (2), anaphora (11) and intensiveness (6).

A specific case belonging to this pattern is the absence of determinant in L, but its presence as demonstrative in one or more translations:

Table 15
– Correspondence in the function of specification

	L	I	F	E1	C	E2	P
Absence	17 (100)%	6 (35%)	1 (6%)	4 (24%)	3 (18%)	2 (12%)	3 (18%)
CD	−	2 (12%)	4 (24%)	8 (47%)	6 (35%)	3 (18%)	3 (18%)
NCD	−	−	−	−	−	−	1 (6%)23
Article	−	9 (53%)	8 (47%)	5 (29%)	6 (35%)	10 (59%)	8 (47%)
Other determinants	−	−	2 (12%)	−	2 (12%)	2 (12%)	2 (12%)
Gap	−	−	2 (12%)	−	−	−	−

• Source: Author’s elaboration. ²⁷

In Table 15, it is seen that the 17 occurrencesof L without determinant correspond preferably to structures with demonstratives (CD and/or DNC) only in E1, being more frequent with articles in I, F, E2 and P and drawing the two cases in C. Anyway, there is an opposition between the prevalence of absence of determinant in L in relation to the different options of formal expression in Romance languages. See below a group of data illustrating the different types of correspondence:

Exc. 22

L: coronatur non Ø coronis que sunt in lege iustorum (f. 2r8) [Absent]

I: sara coronato non de le corone ke sono nela legge de giusti (f. 12r19-20) [Article]

F: elle est courounee. nemie seulement des courounes qui sont en la loi de iustice (f. 309v 6-8) [Article]

E1: es coronado. no de aquellas coronas que son en la ley de los justos (f. 3v10-11) [CD]

C: es coronat. No solament de les corones qui son en la lig dels Justs (f. 1rb32-1va1) [Article]

E2: es coronado no solamente de las coronas que son en la ley de los iustos (f. cxxiij-r27-28) [Article]

P: seeras coroado nõ tam solamente das coroas que som ẽna ley dos Justos (f. 5v4-5) [Article]

The fact that there is often absence of resource in Latin suggests that the perception of the specificity of the referent would be held by discourse-pragmatic means (internal and/or external context) and not by lexical means (with a demonstrative, for example).

In most of the other cases of pairing with gap, there is a gap precisely in the data of L (36 cases), which precludes useful generalizations about the process of change in the demonstrative system from Latin to the Romance languages.

Discussion

Based on the gathered data, it was possible to confirm the hypothesis that the demonstrative systems would have been restructured in the process of language change from Latin to the Romance languages, not just in terms of forms but especially in terms of functions.

Descriptions of the restructuration process of the demonstratives system in terms of forms have been presented in various works of Romance linguistics and we present here a synthesis. According to Väänänen (1995, p.212-213), the demonstrative system was restructured basically by the following changes: (a) is disappeared due to its (formal) brevity, initially taking its place hic; (B) hic disappeared then, passing to play its function iste; (C) ipse lost its own value, came into competition with iste/ille and took the place of idem, occupying in the end the place of iste (originally linked to the second person). The widespread use of the demonstratives in substantive and adjectival function led to the loss of its original value and generated two major change processes involving essentially the demonstrative of 3rd person ille: the creation of a personal pronoun of the 3rd person, from its use in substantive function, and the creation of the definite article, from its use in adjectival function. Väänänen (1995) also notes that in the language used in conversations the demonstratives iste and ille were reinforced with the preceding or prefixed particle ecce (and eccum). Wartburg (1975, p. 138) outlines the restructuring of the demonstratives system as follows:

Figure 2
– Restructuring of the demonstratives system

Source: Wartburg (1975, p.138).

The stages of the restructuring are: I, hic takes the place of is; II, iste extends its scope to the 1st person; III, ipse takes the place of the 2nd person (only in parts of Romania); IV, ille takes the function of personal pronoun of the 3rd person; V, ille or ipse (depending on the region) acquires the function of article; VI, new structures are formed for the expression of reinforcement; and VII, a form of reinforcement (eccille or eccuille) takes the place of demonstrative of 3rd person.

Maurer Jr. (1959), when dealing with demonstratives in Vulgar Latin, draws attention to three important facts:

1. Preserved demonstratives: of the six demonstratives of Classical Latin (hic, iste, ille, is, ipse and idem), only three would be conserved in Vulgar Latin (iste, for 1st person; ipse for 2nd; and ille, for the 3rd in its reinforced form).

2. Reinforcement of demonstratives: in Vulgar Latin the demonstratives were often reinforced with ecce, eccu(m) [< ecce + (h)un(c)] or *accu [< atque or atque + eccu].

3. Vulgar forms and their declination: ipse and ille followed the declination of iste, already reduced to nominative, accusative and genitive-dative.

What descriptions like these (WARTBURG, 1975; MAURER Jr., 1959) do not show is the fact that there were also changes in terms of the recruitment of demonstratives for the expression of certain functions.

As seen in the previous section of data description, the demonstratives would be less frequent in Latin than in Romance languages: this is due, according to what was demonstrated here, to the recruitment of demonstratives to express functions which in Latin were expressed by conjunction (inter-clausal articulator of coordinates and/or adverbial clauses), explanatory phrase (introducer of clarification), present participle (agentive), relative pronoun (inter-clausal articulator of substantive clauses), anaphoric and identity pronoun (anaphora), intensive pronoun (intensiveness) or absence of resource (specification).

It is curious to note that in many cases the change seems to confirm the trend of replacing a synthetic pattern for an analytic one in formation of Romance languages: this applies to the issue of conjunctions (replacement of a conjunctive form by a phrase, as, for example, in the case of ut in Latin by per ço que in Catalan), of present participle (replacement of a verbal-nominal form by a phrase as, for example, in case of volentibus in Latin by coloro che uoglion in Italian), of relative pronouns (for replacement of pronominal form by a phrase, as, for example, in the case of qui in Latin by aquel in Portuguese) and of intensive pronouns (replacement of pronominal form by a phrase, as, e. g., in case of ipsi in Latin by ellos mesmos in Spanish). The formal reduction would lead to the disappearance of a form and the speakers would then recruit another structure, more complex, discursively related to the function expressed by the form that has disappeared, to take the place of the latter.

The issue is, however, still more curious, given that, in some cases, such as the present participle and the relative pronoun, there was a historical continuation of these resources, since the present participle still exists in the Romance languages, as a noun form (see amante in Portuguese), and also as a relative pronoun (see quem in Portuguese, derived from the accusative of qui in Latin). Thus, it is not a matter of filling the gaps, since there is also a process of change in the preference of structures to express certain functions: Catalan, Spanish and Portuguese have relative pronoun to initiate subjective clause (respectively qui, quien and quem), but, in the data (Table 11), the preferred form to represent Latin qui was the phrase with demonstrative (respectively, cell/aquell qui, aquel que and aquele que). One possible explanation for these preferences would be the principle of isomorphism (a form for a function): a forma as qui in Catalan, being homonymous between relative pronoun and interrogative pronoun, would be recruited preferentially to one of these functions (in this case, the interrogative), being recruited to another function (the relative one) the phrase with the demonstrative. Complementary to the principle of isomorphism in this particular case, there would be the need for disambiguation: since qui in Catalan, for example, is interrogative and relative, there could be contexts in which it would be difficult to realize which function it would be expressing and this would accentuate a previous trend to separate a form for each function in this case.

To rigorously prove the action of the principle of isomorphism in the selection of demonstratives a much more extensive work will be needed, identifying all the functions that the demonstratives express and their competing forms. The data gathered here have made it clear that the selection of demonstratives is, in most cases, a trend, not a categorical process, which means having to insert, in more extensive analysis, a variationist perspective.

One last point is related to the methodological question of this study: the study of translations has any bias that can determine certain patterns in the data?

A first answer is yes: just check the patterns between E2 and P to verify the fact that the Portuguese translation being derived from the Spanish one is a reason for the similarity in the distribution of the data. As an example, the data in Table 12, where E1 and E2 are radically different (although they are the same language, Spanish), but E2 is fully in line with C (E2 derives from a Catalan model) and with P (E2 represents a tradition that served as a model to P). One must consider, however, that E1 and E2 may reflect dialectal differences: E1 would have been translated by Catalan Father Bernardo Boil (c.1445-c.1520), who even says that his translation is in Aragonese²⁴, while E2, having been published in Seville, may reflect the dialect of this other region. In a previous study on E1 considering five phonological aspects (CAMBRAIA, 2007), however, there was no detection of any proper Aragonese characteristic, only of Spanish in general.

It should be noted, however, that the subordination of one translation to another did not mean absolute obedience to the pattern of the model: see, e. g., in Table 9, that E2 and P use preferably CD as a resource for conjunctive phrases, but not to the same extent, indicating that the translator of P was able to detach from his Spanish model, to be, perhaps, more faithful to his own linguistic pattern.

Final considerations

The study carried out here showed that, alongside formal changes in Latin demonstrative system to Romance languages, there was also functional changes, with the increasing recruitment of demonstratives to express functions which, in Latin, were expressed by other resources.

The validation of the hypothesis of the functional changes of demonstratives suggests that, for new studies, one should prioritize the functional aspect in the description of these elements in order to identify not only the functions they express but also the forms that compete with the demonstratives in expressing such functions, opening therefore space for the inclusion of the variationist perspective in the discussion about the history of this word class.

It will also be interesting doing the same study (comparative, empirical and functional) with data of modern Romance languages to evaluate whether there has been a simple continuation of the medieval trends of preference for demonstratives in certain functions or whether these functions would have come to be expressed by new forms.

REFERÊNCIAS

• AVELLAR, J. Tradição latino-românica do Livro de Isaac: subsídios para uma edição crítica da tradução catalã. 2011. Relatório (Iniciação Científica) ‒ Faculdade de Letras, Universidade Federal de Minas Gerais, Belo Horizonte, 2011.
• BADÍA I MARGARIT, A. M. Gramàtica històrica catalana 3. ed. València: Tres i Quatre, 1994.
• BRUGMANN, K. Die Demonstrativpronomina der indogermanischen Sprachen: eine bedeutungsgeschichtliche Untersuchung. Abhandlungen der Sächsischen Gesellschaft der Wissenschaften, Leipzig, n. 22, p. 1-50, 1904.
• BÜHLER, K. Sprachtheorie: die Darstellungsfunktion der Sprache. Stuttgart: G. Fischer, 1934.
• CAMBRAIA, C. N. Dulceça, dulçor, dulçura e dulcidom: um estudo de caso de variantes derivacionais no português medieval. Estudos de Lingüística Galega, Santiago de Compostela, v.2, p.37-56, 2010. Disponível em: <http://www.usc.es/revistas/index.php/elg/article/download/1507/1369> Acesso em: 28 mar. 2016.
  » http://www.usc.es/revistas/index.php/elg/article/download/1507/1369>
• CAMBRAIA, C. N. Demonstrativos nas traduções medievais espanholas do Libro Llamado Abbat Ysach. In: CONGRESSO BRASILEIRO DE HISPANISTAS, 5.; CONGRESSO INTERNACIONAL DA ASSOCIAÇÃO BRASILEIRA DE HISPANISTAS, 1., 2008, Belo Horizonte. Anais... Belo Horizonte: Faculdade de Letras da UFMG, 2009. p. 2380-2388. 1 CD-ROM. Disponível em: <http://150.164.100.248/espanhol/Anais/anais_paginas_%202010-2501/Demonstrativos%20nas%20tradu%E7%F5es.pdf> Acesso em: 28 mar. 2016.
  » http://150.164.100.248/espanhol/Anais/anais_paginas_%202010-2501/Demonstrativos%20nas%20tradu%E7%F5es.pdf>
• CAMBRAIA, C. N. Libro llamado Abbat Ysach (1489): aragonês e/ou castelhano?. Belo Horizonte, 2007. Comunicação apresentada no V Congresso Internacional da Abralin, realizado na Universidade Federal de Minas Gerais, em Belo Horizonte, no período de 28 de fevereiro a 3 de março de 2007.
• CAMBRAIA, C. N. Introdução à crítica textual. São Paulo: Martins Fontes, 2005.
• CAMBRAIA, C. N. A difusão da obra de Isaac de Nínive em línguas ibero-românicas: breve notícia das tradições portuguesa, espanhola e catalã. In: RAVETTI, G.; ARBEX, M. (Org.). Performance, exílio, fronteiras: errâncias territoriais e textuais. Belo Horizonte: Ed. da UFMG, 2002. Disponível em: <http://150.164.100.248/profs/cesarnardelli/data1/arquivos/2002-Cambraia-Difusao.pdf> Acesso em: 28 mar. 2016.
  » http://150.164.100.248/profs/cesarnardelli/data1/arquivos/2002-Cambraia-Difusao.pdf>
• CAMBRAIA, C. N. Livro de Isaac: edição e glossário. 2000. 753f. Tese (Doutorado em Filologia e Língua Portuguesa) ‒ Faculdade de Filosofia, Letras e Ciências Humanas, Universidade de São Paulo, São Paulo, 2000. Disponível em: <http://150.164.100.248/profs/cesarnardelli/data1/arquivos/2000-Cambraia-Isaac.pdf> Acesso em: 28 mar. 2016.
  » http://150.164.100.248/profs/cesarnardelli/data1/arquivos/2000-Cambraia-Isaac.pdf>
• CAMBRAIA, C. N.; BIANCHET, S. M. G. Caleidoscópio latino-românico: demonstrativos. Cadernos de Letras da UFF, Niterói, v. 35, p.15-35, 2008. Disponível em: <http://www.uff.br/cadernosdeletrasuff/35/artigo1.pdf> Acesso em: 28 mar. 2016.
  » http://www.uff.br/cadernosdeletrasuff/35/artigo1.pdf>
• CAMBRAIA, C. N.; LARANJEIRA, M. B . Tipologia dos erros na tradição latina do Livro de Isaac. Caligrama: revista de estudos românicos, Belo Horizonte, v. 15, p. 7-48, 2010. Disponível em: <http://www.periodicos.letras.ufmg.br/index.php/caligrama/article/view/26/28> Acesso em: 28 mar. 2016.
  » http://www.periodicos.letras.ufmg.br/index.php/caligrama/article/view/26/28>
• CAMBRAIA, C. N.; CUNHA, E. L. P. T. Tradição em língua catalã do Livro de Isaac. Scripta Philologica, Feira de Santana, n. 4, p. 119-167, 2008.
• CAMBRAIA, C. N.; MELO, T. C. A de; VILAÇA, C. E. de L. Tradição latino-romanica do Livro de Isaac: analise de alguns lugares-críticos. Filologia e Linguística Portuguesa, São Paulo, v. 10/11, p. 409-425, 2008/2009. Disponível em: <http://www.revistas.usp.br/flp/article/view/59832/62941> Acesso em: 28 mar. 2016.
  » http://www.revistas.usp.br/flp/article/view/59832/62941>
• CAMBRAIA, C. N.; VILAÇA, C. E. de L; MELO, T. C. A. de. Unidade lexical e unidade cultural: o léxico românico de religião em traduções medievais. Traduções, Florianópolis, v. 5, p. 22-39, 2013. Disponível em <http://www.incubadora.ufsc.br/index.php/intraducoes/article/view/2557/3233> Acesso em: 28 mar. 2016.
  » http://www.incubadora.ufsc.br/index.php/intraducoes/article/view/2557/3233>
• DEES, A. Étude sur l’évolution des démonstratifs en ancien et en moyen français Groningen: Wolters-Noordhoff, 1971.
• DIESSEL, H. Demonstratives: form, function and grammaticalization. Amsterdam: Benjamins, 1999.
• DIXON, R.M.W. Demonstratives: a cross-linguistic typology. Studies in Language, Amsterdam, v.27, n.1, p.61-112, 2003.
• FARIA, E, Gramática superior da língua latina Rio de Janeiro: Acadêmica, 1958.
• FRANÇA, C.S. Edição e estudo linguístico das traduções em línguas ibero-românicas do tratado ascético-medieval Livro de Isaac: subsídios para o estudo da tradução espanhola. 2004. Relatório (Iniciação Científica) ‒ Faculdade de Letras, Universidade Federal de Minas Gerais, Belo Horizonte, 2004.
• GIVÓN, T. Syntax: an introduction. Amsterdam; Philadelphia: John Benjamins, 2001. 2 v.
• GIVÓN, T. Functionalism and grammar Amsterdam; Philadelphia: John Benjamins, 1995.
• GODEFROY, F. Dictionnaire de l’ancienne langue française et de tous ses dialectes du IX^e au XV^e siècle. Paris: F. Vieweg, 1885.
• KELLER, R. M. Iste deiktikon in the early roman dramatists. Transactions and Proceedings of the American Philological Association, v. 77, p. 261-316, 1946.
• LAUSBERG, H. Lingüística románica 2. ed.Lisboa: Calouste Gulbenkian, 1981.
• MARCHELLO-NIZIA, C.; PICOCHE, J. Histoire de la langue française Paris: Nathan Université, 1998.
• MARTELOTTA, M. E.; AREAS, E. K. A visão funcionalista da linguagem no século XX. In: CUNHA, M. A. F. da; OLIVEIRA, M. R. de; MARTELOTTA, M. E. Lingüística funcional: teoria e prática. Rio de Janeiro: FAPERJ: DP&A, 2003. p.17-28.
• MAURER JR., T. H. Gramática do latim vulgar Rio de Janeiro: Acadêmica, 1959.
• MAURER JR., T. H. A unidade da România ocidental São Paulo: Edusp, 1951.
• MELO, T. C. A. Livre d’Isaac Abbé de Syrie (cód. lat. 14891 da BNF): edição e glossário.2010. 370f. Tese (Doutorado em Estudos Linguísticos) ‒ Faculdade de Letras, Universidade Federal de Minas Gerais, Belo Horizonte, 2010. Disponível em: <http://150.164.100.248/poslin/defesas/1233D.pdf> Acesso em: 28 mar. 2016.
  » http://150.164.100.248/poslin/defesas/1233D.pdf>
• NEVES, M. H. de M. A gramática funcional São Paulo: Martins Fontes, 1997.
• OUY, G. Les manuscrits de l’Abbaye de Saint Victor: catalogue établi sur la base du repertoire de Claude Grandrue (1514). Turnhout: Brepols, 1999. 2 tomos.
• VÄÄNÄNEN, V. Introducción al latín vulgar 3. ed. 1. reimpr. Madrid: Gredos, 1995.
• VILAÇA, C. E. de L. Libro dell´Abate Isaac di Siria: edição crítica e glossário. 2012. 345f. Tese (Doutorado em Estudos Linguísticos) ‒ Faculdade de Letras, Universidade Federal de Minas Gerais, Belo Horizonte, 2012. Disponível em: <http://150.164.100.248/poslin/defesas/1152D.pdf> Acesso em: 28 mar. 2016.
  » http://150.164.100.248/poslin/defesas/1152D.pdf>
• VILAÇA, C. E. de L. Libro dell´Abate Isaac di Siria (cód. ricc. 1489 da BRF): edição e confronto com a edição princeps de 1500. 2008. 418 f. Dissertação (Mestrado em Estudos Linguísticos) ‒ Faculdade de Letras, Universidade Federal de Minas Gerais, Belo Horizonte, 2008. Disponível em: <http://150.164.100.248/poslin/defesas/1152M.pdf> Acesso em: 28 mar. 2016.
  » http://150.164.100.248/poslin/defesas/1152M.pdf>
• VOTRE, S.; NARO, A. J. Mecanismos funcionais do uso da língua. In: MACEDO, A. T. de; RONCARATI, C.; MOLLICA, M. C. (Org.). Variação e discurso Rio de Janeiro: Tempo Brasileiro, 1996. p.51-62
• WARTBURG, W. von. Problemas e métodos da linguística São Paulo: Difel, 1975.

• 1
  In the original: “A hipótese fundamental desta proposta é que do uso da língua — a comunicação na situação social — origina-se a forma da língua, com as características que lhe são peculiares, inclusive, diferentes graus de instabilidade associados a diferentes subsistemas. Isso supõe entender a língua como um objeto maleável, probabilístico e não-determinístico. Portanto, nessa visão, a estrutura (ou forma da língua) é uma variável dependente, resultante de regularidades das situações em que se fala.”
• 2
  In the grammatical tradition, it is common to name these types as adjective and pronoun.
• 3
  In the original: “Não negamos a existência de estrutura, pelo contrário queremos entender as suas motivações básicas, admitindo que essas podem ser exclusivamente diacrônicas em determinadas situações. Entretanto, julgamos oportuno observar que alguns funcionalistas não compartilham de nossa posição moderada, considerando [os referidos funcionalistas] que a estrutura não possui existência independente do uso da língua [...].”
• 4
  We present the data of C, E2 and P always in this sequence, in order to facilitate the perception of the influence of the model of one over the other, since, as it will be explained later, the translational route must have been Catalan > Spanish (represented by E2) > Portuguese.
• 5
  For further discussion of the genetic relationship between the testimonies, see Cambraia (2002, 2010, 2010) and Cambraia, Melo and Vilaça (2008/2009).
• 6
  Just recall the thesis of Maurer Jr. (1951) that the unity of western Romania is related, among other things, to the learned influence of Latin. This influence was in fact found in a study of the lexicon of religion in Latin, Italian, French and Portuguese translations of the work of Isaac which are the subject of this study (CAMBRAIA; MELO; VILAÇA, 2013).
• 7
  The transcriptions were adapted from the following sources: L, Cambraia and Laranjeira (2010); I, Vilaça (2008); F, Melo (2010); E1 e E2, França (2004); C, Avellar (2011). The one of P was made directly from the manuscript. Due to the adoption of short reproductions in the examples, we chose to remove all punctuation, since it wouldn´t contribute to the understanding of the examples.
• 8
  For ease of general reference to these forms, we use the following abbreviations: in ternary systems (L, I, E1, C, E2 and P) F1 for the first of the serie; F2, for the second; and F3 for the third; in binary systems (F), FI, for the first of the serie; and FII for the second.
• 9
  Abbreviations: nm. = nominative; gn. = genitive; ac. = accusative; dt. = dative; ab. = ablative.
• 10
  Determinants (22%): hac₁ (f.s.) [ab.], hoc₂ (m.s.) [ab.], huius₅ (m.s.) [gn.], illa₁ (f.s.) [ab.], illo₁ (n.s.) [ab], illam₂ (f.s.) [ac].
• 11
  Throughout this text, we use italics in the examples to identify the forms in discussion.
• 12
  Determinants (33%): questo₅, questa₂, queste₆, quel₂, quella₅, quelli₁, quelle₆. It is worth noting here the form quell, that only appears as a determinant in the corpus: the absence of final vowel (vs. quello) is due precisely to its status as proclitic form, therefore dependent on a nucleus.
• 13
  Abbreviations: NC = Nominative Case; DOC = Direct Oblique Case (non prepositional); IOC = Indirect Oblique Case (prepositional).
• 14
  Determinants (38%): este₈, esta₇, estas₆, essa₂, aquel₂, aquella₅, aquellos₁, aquellas₉.
• 15
  Determinants (30%): aquest₆, esta₁, aquesta₂, aquestes₁₂, aquell₁, aquella₅, aquelles₄. These data allow verifying, in relation do Table 5, that cell and cells are forms reserved to be used as independent pronouns.
• 16
  Determinants (18%): este₂, aqueste₁, esta₃, estas₉, aquel₃, aquella₄, aquellas₃.
• 17
  Determinants (22%): este₅, esta₄, estas₉, aquel₃, aquela₃, aquella₁, aquellas₂. It is interesting to notice here that, differently of Italian, the form without final vowel aquel doesn’t occur only as determinant in medieval Portuguese.
• 18
  In an extensive study on the CDs of the two Spanish translations (CAMBRAIA, 2008, p. 2381-2382), the following values were verified: E1, F1 670 (58%), F2 47 (4%), F3 439 (38%), total 1.156 (100%); E2, F1 728 (58%), F2 20 (2%), F3 494 (40%) and total 1.242 (100%). The total data situate, as expected, the ternary system of E1 in the pattern of preference for the forms of proximity and/or 1 person (F1). One can see that the selected corpus comprises approx. 11% of the total forms in each Spanish translation.
• 19
  We use the sign Ø to mark the absence of corresponding form or passage in relation to the other translations.
• 20
  We don´t inform the folios of each record in this table in order to avoid information overload and consequent reading difficulty.
• 21
  We estimate, based on some Romance translations, that the genuine form was volentibus (“who want”), since the idea of denial would be expressed in conjunction ne.
• 22
  Although esse derives historically from ipsae (intensive pronoun), it is considered to be a personal pronoun of 3rd person (PP3) already in medieval Italian, having taken the function of intensive pronoun the forms stesso (< istu ipsu) or medesimo (< metipsissimu).
• 23
  In this case, the NCD isn’t followed by a noun, rather introduces a relative clause: “cõprender o que ha de uijr por conhecimento dos entendimẽtos”.
• 24
  “Pedistes me senyor enlos dias passados: el nuestro Abbat ysach el qual yo por su marauillosa doctrina: y ensenyança / a ruego delos padres / y hermanos desta nuestra montanya en el comienço de mi conuersion de latino hauia fecho Aragones / o si mas querres Castellano. no daquel mas apurado stilo dela corte. mas daquel llano que ala profession nuestra segun la gente / y tierra donde moramos paraque le entiendan satisfaze.” (f. Aij-v, l. 1-12; italic added).

Publication Dates

• Publication in this collection
  Jan-Apr 2016

History

• Received
  Feb 2015
• Accepted
  Apr 2015