Clitic Placement and the Grammaticalization of the Future and the Conditional in Old Catalan

Round 1

Reviewer 1 Report

The topic is interesting and methodologically well framed, but would benefit from broader contextualization and exemplification from the perspective of Romance and non-Romance languages alike. The article additionally has the potential to appeal to an audience versed in sociolinguistics, a path that the author may consider exploiting further in reframing their rationale for the present contribution. It would be useful for readers to be forewarned on the controversy still surrounding the topic of grammaticalization, with the author clearly stating their particular stance on the matter (Hopper & Traugott's concept of grammaticalization paths or 'clines' may be helpful in this respect--and so might be illustrating how grammaticalization tends to perform as a 'work-in-progress' kind of phenomenon). The article is well written, and attention has been paid to language accuracy in terms of grammatical structures and wording choices. A number of sentences and clauses appear to bear a Romance-like academic style, which may occasionally not work for mainstream academic English (e.g., "With this aim, a diachronic corpus has been made up" would flow better if reframed along the lines of "A diachronic corpus was compiled [in order] to serve this purpose"). A suggestion is also made that the author revise the appropriateness of Greek- and Latin-based technical terms like 'mesoclisis' or 'univerbiation,' since those do not always translate well into English (at any rate, readers might benefit from a brief definition of those terms prior to their effective use). 'Old Catalan' is understood on pair with the formulaic use as applied to Germanic languages; still, the author might want to consider 'medieval Catalan,' for the sake of preciseness and consistency with the standards for the Romance subfield. A list of additional, relatively recent (for the most part) publications is provided below:

Alturo, N. (2017). L’auxiliar de passat ‘anar’ en català. Apunts de pragmàtica del català. Teories de la gramaticalització: Cas 1. Barcelona: Dipòsit Digital de la Universitat de Barcelona, 2-9.

Alturo, N., & Chodorowska-Pilch, M. (2009). La gramaticalització de “sisplau”. Els Marges, 88, pp. 15-38.

Bisang, W. (2017). Grammaticalization. Oxford Research Encyclopedia of Linguistics.

Cuenca, M. J. (2001). La gramaticalització com a teoria de la variació morfosintàctica. La gramàtica i la semàntica en l’estudi de la variació, pp. 151-185. Barcelona: PPU. 

Cuenca, M. J., & Massip, À. (2005). Connectors i processos de gramaticalització. Caplletra, 38, pp. 77-92.

Gandarillas, M. (2021). The grammaticalization of Catalan anar (‘to go’) + infinitive for the expression of perfective past: A diachronic, corpus-based perspective. International Linguistics Research, 4(4).

Gandarillas, M. (2020). Grammaticalization degrees in Catalan anar vs. estar + adjective in the 19^th and 20^th centuries: A language contact, corpus-based distributional approach. Isogloss: Open Journal of Romance Linguistics, 6, 1–18.

Gandarillas, M. (2020). El papel de la subjetivación en los procesos de gramaticalización. Spanish and Portuguese Review, 6, 73–86.

Juge, M. L. (2005). Morphological factors in the grammaticalization of the Catalan “go” → past. Diachronica, 23(2), pp. 313-339. 

Montserrat i Buendia, S. (2004), Entre la dixi i la definitud: Els verbs de moviment resultatiu en català. Caplletra, 39, pp. 61-83.

Norde, M. (2019). Grammaticalization in morphology. Oxford Research Encyclopedia of Linguistics, vol. Morphology.

Pérez-Saldanya, M., & Hualde, J. I. (2003). On the origin and evolution of the Catalan periphrastic preterit. Verbalperiphrasen in den (ibero)romanischen Sprachen. Hamburg: Buske. 46-70.

Traugott, E. C. (1995). Subjectification in grammaticalisation. Subjectivity and Subjectivization: Linguistic Perspectives, pp. 31-54. Cambridge: Cambridge University Press.

Author Response

Thank you very much for your comments and suggestions. To improve the paper, we have broadened the state of the art and we have also included a more general Romance perspective. As you mention, we would like to focus on the sociolinguistic aspect of this phenomenon in future works. The paper was already proof reviewed by an English native speaker, but we have revised some language aspects once again and we have followed the linguistic corrections that you and the other reviewer have suggested. Thank you very much for noting them. We have reviewed the terms ‘mesoclisis’ and ‘univerbation’ and we have checked that both are used in English papers on this topic. Therefore, we consider that they are sufficiently well-known in the previous bibliography. Thank you for letting us know your concerns. Likewise, ‘Old Catalan’ is the usual term to refer to the period under study in this paper. Nevertheless, the chronology is very detailed in the text, so we do not think it could cause any problem.

Reviewer 2 Report

The paper has a preeminently descriptive and empirical (quantitative) focus and is of considerable import on this account, since it constitutes the first exhaustive depiction of the relative frequencies of the different possibilities for clitic placement (proclisis, mesoclisis, enclisis) associated with Catalan future/conditional constructions throughout the Middle Ages and the 16^th c., thus decisively contributing to complete a panorama of Ibero-romance varieties where Navarrese-Aragonese and Castilian have been thoroughly studied for the same phenomena over the last few years, allowing the author of this paper significant comparisons as to the placement of Catalan within the Romance continuum. As it turns out, Catalan privileges enclisis to a root+inflection base far more than Navarrese-Aragonese, while in central and western Castile and in Portugal this phenomenon is practically unknown until well into the 15^th c. Furthermore, the advance of proclisis in contexts that could alternate this ordering with the other string alignments, and even in contexts where proclisis was originally banned, clearly appears to have advanced more speedily in Catalan (with very significant progress already in the 14^th c.) than in any other variety (similar cases of proclitic advancement are still rare in Castilian by 1500). The author proves these significant quantitative differences beyond any possible doubt through a detailed analysis of all data in a well-balanced corpus of selected texts from the most encompassing and reliable Catalan historical digital corpus (CICA), showing a masterful ability to interpret big sets of data while not shying away from in-depth philological detail (suffice the remarks on a particular text, the Vides de sants rosselloneses, that displays even more proclisis than other texts in Catalan due to its origin in Rousillon, already close to Provence, where the tendence to abandon other orderings is even greater). The author also makes an important connection with the greater inclination of Catalan to display “altered” root morphology (shortened, syncopated, sometimes later undergoing plosive insertion), which she rightly interprets as a sign of greater coalescence within the bonded root+inflection futures/conditionals.

All in all, the paper hints at a Romance continuum regarding this phenomenon that runs approximately from Northeast to Southwest, with langue d’oil refusing mesoclisis and enclisis, Old Italian allowing enclisis but not mesoclisis, langue d’oc allowing all three alignments but clearly favouring proclisis, and Catalan progressively tending toward proclisis (while keeping the balance between mesoclisis and enclisis) at a visibly faster rate and in earlier stages than neighbouring Aragonese, which however still admits enclisis, whereas Castilian is very reluctant to the clitic-final configuration in the Middle Ages, and Portuguese does not seem to admit it and (at least in its European variety) displays mesoclisis up to this day. The more western the variety, the less rapidly mesoclisis is lost and proclisis advances, and the more transparent future/conditional roots are. There is thus a correlation between the wider array of alignments, the chronology and velocity of proclitic expansion, and the transparency of the (originally) infinitival part of the compound (the author follows Alsina 2020 in considering that, at least at the interface with phonology, all three configurations can be handled as words, with mesoclitic strings as compound words). This is sufficiently demonstrated by the author and taken as evidence that the same continuum can be interpreted in terms of a greater degree of formal grammaticalisation of the futures/conditionals in the “eastern” varieties, with Catalan, as it were, in a highly interesting fulcrum position.

This is, in sum, a well-documented, convincingly argued paper that in my view deserves publication without major modifications. I do, however, have suggestions for minor modifications which I propose in the following.

My greater concerns regard the way in which former research on this topic (not particularly on Catalan, which had not been much dealt with until Alsina 2020, but rather on Castilian) is presented. The author opposes what she clearly regards as the mainstream hypothesis (as postulated by Castillo or Bouzouita), that has “proved that clitic placement in the FC is governed by the same syntactic-pragmatic restrictions as other tenses” (p. 2), and Company’s idea that the so-called “analytic” FCs are “marked construction[s] with a modal meaning and with a pragmatic function” (topicalization strategy)” (ibid.). This state of the art is rather unsatisfactory. On one hand, it misses out on a series of other analyses that have insisted on the fact the infinitive in the “analytical” FCs are fossilised constructions that result from an alternative grammaticalisation path starting in Late Latin (Girón Alconchel 2007) and where the infinitive has an information-structural or modal value, rather one assignable to the Force node (Batllori 2016) or expressing weak focus (Octavio de Toledo 2015) than a topicalising value, pace Company. On the other hand, the author does not seem to realise that both claims (Castillo/Bouzouita’s and Company and others’) are not incompatible: it is fine to assert that the clitic configurations are the same as with other tenses (and other periphrastic constructs, in the case of the “analytical” FC), but the question raises immediately what kind of properties regulated the selection of the corresponding configuration. In the more traditional view (as instantiated in modern times by Castillo and Bouzouita’s first works on the topic), it is just properties at the phonological level that were (gradually) being lost by the end of the Old Castilian period (the 15^th c.). In the view of Company, Octavio de Toledo, Batllori and Bouzouita’s latest works, the properties at stake concern word order structure, i.e. in the range of possibilities for fronting (whether of a finite or infinitival form) in a given a variety (Medieval Castilian, in this case) and the historical modifications whereby those fronting possibilities were lost. An interesting insight that is gained from the latter perspective is that mesoclisis and enclisis are basically the same configuration, since enclisis has a clitic attached to a fronted finite form and mesoclisis has a clitic attached to a fronted infinitive. This helps to explain why, in varieties like Catalan where enclisis and mesoclisis coexist, these do not necessarily compete among them, but rather align as a block versus proclisis, which corresponds to a syntactic setting where the old fronting properties do not longer apply. If this or a similar line of analysis is not applied, and everything there is consists on a substitution of a less grammaticalised form (in terms of coalescence) by a more grammaticalised alternative, one is left out to wonder why the already coalesced enclitic alternative did not take over the “analytic” FC (the author’s data show stability over time as regards the mesoclisis/enclisis alternation), and also why proclisis progressively advanced into the terrain of enclisis in all varieties if the former configuration already “solved the problem” posed by analytic FCs, i.e. their “insufficient” grammaticalisation. The processes so well described by the author in this paper do not seem to portray a “therapeutical” or semi-automatic change whereby unsufficiently coalesced forms are replaced by coalesced ones just for the sake of greater grammaticalisation (it is, by the way, dubious that greater grammaticalisation always correlates with an increase in frequencies, pace Bybee [cf. p. 6 in this manuscript]: as she herself puts it, “[a]s long as frequency is on the rise, changes will move in a consistent direction […]. When a grammaticalization construction ceases to rise in frequency, various things happen, but none of them is the precise reverse  of the process” (Bybee 2011: 77) – which implies that grammaticalising constructions may lose frequency, although this is not to be seen as degrammaticalisation). The author does not need to provide an answer to these questions here, since it would imply a whole new investigation, but in my opinion he does need to acknowledge that these questions persist, and that he does not provide an answer to them, nor do I believe that in some future research she will prove “that the AFC do not perform any special pragmatic function”, not least because, from a methodological point of view, pragmatic associations and values can be suggested on syntactic or contextual proof, but never strictly ruled out (to put but one similar example, we may not have found yet any particular proof of individual pragmatic properties regulating the presence of an overt pronominal subject vs. its absence in Old Romance, but we cannot rule out that pragmatic properties were involved in that alternation – as a matter of fact, most historical linguists rather tend to assume they were, somehow). But, most importantly, the author simultaneously defends the idea that no pragmatic import is at stake for analytic FCS while at same time somewhat contradictorily assuming (p. 3, lines 84-85) that there were fronted participles in Old Romance and that “emphasis accounts for the presence of preverbal clitics”, i.e. that both infinitive and clitic place could be regulated by information-structural or pragmatic rules, apparently except for the old FCA, which would be somehow immune to such principles. In order to confront these remarks properly, I believe the author should start by tracing a fairer, better-balanced state of the art along the suggested lines.

While very interesting and, in my opinion, quite right, the correlation between a greater frequency of root “alterations” in the FCs (syncopes, epentheses, etc.) and a lesser frequency of “analytic” FCs was already pointed out for Castilian by Moreno Bernal (2004), a publication which in my view deserves a place in the bibliography as an antecedent of the idea. As for the greater frequency of such alterations in the Eastern half of the Peninsula, Saralegui (1985) already identified considerable differences in this respect when comparing Castilian and Navarrese-Aragonese Medieval texts.

Regarding theoretical terminology, I think the term “emphasis”, profusely advocated through the paper, is way too vague to constitute an explanatory device, and one may suspect it pretty much amounts to a synonym of ‘some strange thing we do not quite know’ (cf. in this respect criticism of the notion by López Serena 2012). At the very least, the use of the notion here should be upholstered by referring to those pragmatists that have employed it to describe semantic values in certain pragmatically sensible configurations (e.g. Leonetti/Scandell 2021), i.e. the notion should be somehow justified, anchored in a theoretical frame, and restrained to render it meaningful.

The author also risks the general statement that the reduced forms of former HABEO are “no longer a proper auxiliary” (fn. 2) in all three Fut/Cond+clitic strings, an idea which, to the best of my understanding, does not directly derive from Alsina (2020), quoted as source in the footnote, since the latter conceives of the “analytical” FC as a (compound) word “consisting of an infinitive with a clitic cluster and a bound auxiliary” (Alsina 2020: 198). Since “bound”, however, refers to morphophonological structure, not syntactic properties, the reduce forms of HABEO can still be regarded in this construction as a proper syntactic auxiliary, irrespective of the way this property is structurally or materially spelled out.

In p. 9 (322-323), the author speaks of “the factors that entailed SFC-p”. Strictly speaking, no factors ‘entailed’ SFC-p, since the other two configurations were also possible (if unlikely at some stages) in the same structural position. It would be best to rephrase as “the factors that favoured SFC-p”. Same applies to the use of ‘entail’ in p. 11, l. 411.

It is entirely unassumable that the enclitic construction “spread from Catalan […] to other languages and varieties in the Iberian Peninsula during [the] 11^th-13^th centuries”. The odds for Catalan-Castilian direct contact during those centuries are simply too unlikely. The enclitic solution was adopted in central and western Castile by the end of the 15^th c., and its appearance in highly elaborated, literary materials suggests that its adoption was driven then by Italian influence and possibly also influenced by Eastern Ibero-Romance (Aragonese, Catalan) similarly elaborated prose. Early appearances of the enclitic string in Eastern Castilian texts (e.g. the E6 Bible or the Cid manuscript) are no doubt due to contact with the immediate Navarrese-Aragonese area, not distant Catalonia. Thus, Catalan may have a cultural impact in the diffusion of the enclitic configuration among literate individuals in Castile starting 1450, but not a direct influence via oral contact among populations. This passage should, in my opinion, be more properly formulated.

The author claims that in a configuration such as defendre-nos hien “the stress remains far from the word ending, in the syllable -fen-, and in Catalan, oxytones and words stressed on the penultimate syllable are far more common” (p. 8). Whilst I understand and agree with the accent-pattern argumentation, it must be remarked that the authors have assumed, with Alsina (2020), that the “analytic” future is a as a whole only one word, with an accent on the infinitive-like formant and another accent on the bound auxiliary. Hence, strictly speaking, this string is oxytonic, as any other AFC. What the authors seem to imply is that it is far more common for infinitive+clitic strings in Catalan to be stressed on the last-but-one syllable.

The factor “verbal tense” as possible variable that favours the appearance of certain strings is invoked a couple of times throughout the paper (e.g. p. 9, l. 314-315), but never discussed or assigned any value. Even though the author does not go deep into the analysis of these variables here, which is perfectly legitimate, he should nevertheless inform the reader what she means by the variables she mentions. I guess the answer here is that both mesoclisis and enclisis are only possible in present or imperfect, a constitutive difference with proclisis, which can affect any verbal tense. If this is so, it should be briefly worded somewhere. A few lines afterwards, “the first conjugation” should be shortly described for all readers unfamiliar with Catalan or Romance in general.

The paper is neatly structured and generally well written, and can thus be easily followed from start to end without difficulty. It is however unidiomatic and somewhat inelegant at some passages, which I have marked blue(ish) in the accompanying file. In particular, unidiomatic verb-object constructions with have (“has two different stages”, “synthetic forms have univerbation”, “The remaining centuries do have the expected development”, instead of ‘shows’, ‘display’ and similar verbs) appear in a number of different places. Similarly, the use of ‘occur’ is strange in “as occurs with” or “unlike that which occurs in Spanish” (both p. 9), where one would expect “as is the case with” or simply “unlike in Spanish”. Instead of “the environments in Group 2 regarding main sentences”, I think “the root sentence environments in Group two” is more correct. Also, “as follows” is (wrongly) used as ‘hereafter’ (p. 6, l. 201). The sentence “literary Catalan, which is deemed unusable with its only 2 tokens” (p. 6, l. 222) clearly deserves better phrasing. And I cannot understand what the authors mean by the following: “, the FC are the only two verbal tenses of the entire verbal conjugation of Catalan” (p. 6, fn. 8) ‒ they probably want to state that those tenses are the only ones derived from a former Latin periphrasis, but if this is so, the sentence must be rephrased accordingly. In that same fn., ‘among’ is preferable to ‘of’ in the following phrase: “the most common construction of the three possible positions of clitics”. Again in fn. 8, “because of its limited extension” should be “because of their limited extension”, since the pronoun seems to retake the word ‘texts’. Finally, I am aware of but a few typographic mistakes, which I have marked red on the file.

Comments for author File: Comments.pdf

Author Response

Thank you very much for this informative review. We really appreciate your comments, suggestions, and argumentation. Undoubtedly, they will improve our paper. Most of your comments have been addressed in our new draft. Some others, as you state in your review, will help us to deal with future investigations.

We have addressed our paper in the following ways:

1. About the state of the art: We are grateful for the discussion that you have provided us. Certainly, the state of the art, too brieve in our first draft, should be improved. We have addressed it following your suggestions, that is, including a more varied bibliography and the explanation of different hypotheses to explain the AFC/SFC-p variation. We have summarized these papers just before our section 1.1. However, we would like to point out that we do not consider compatible all the explanations given by the bibliography, as seems to be mentioned in the revision (“the author does not seem to realise that both claims (Castillo/Bouzouita’s and Company and others’) are not incompatible”). As we have added in the paper, there are some important points that appear to be against the reading of AFC as a marked configuration. Nevertheless, we think that now the state of the art is more well-balanced than before.
2. We have also added Moreno Bernal (2004) and Saralegui (1985). We agree that they should be in the bibliography of this paper.
3. About “emphasis”: We have clarified what we understand for “emphasis”, and that we are following Granberg (1999) and recent bibliography on informative structure (Leonetti and Escandell-Vidal 2021). We would like to refer to the supposed problematic of viewing a pragmatic import in SFC-a while ruling out a pragmatic import in the use of AFC. We do not pretend to declare that the AFC cannot receive pragmatic meanings; what we declare is that pragmatics do not decide which form is used in the AFC/SFC variation, since, first of all, clitic placement has to be taken into account, and, secondly, there are morphological factors (i.e., verbal syncope, addressed in the paper) which explain some of the AFC/SFC-p variation. On the other hand, an extensive bibliography has observed the role of emphasis in the presence of proclitic pronouns, as it is included in the paper. Nevertheless, we are aware that in future investigations we will have to study this phenomenon more in depth. Thank you for noting that.
4. Regarding the diatopic hypotheses, we would like to clarify that we do not suppose a direct influence of Catalan over Castilian. Perhaps, the phrasing of our explanation of this point was not clear enough. We have tried to complete it. The diatopic hypotheses suggests an east-to-west phenomenon: Occitan and Catalan would directly influence on Navarro-Aragonese (Occitan in Navarra, since the Navarrese court used Occitan for some time, and Catalan with Aragonese by linguistic contact in the borders between the two varieties and also because of the politic relationship of the two languages within the Crown of Aragon). As the review says, Castilian has indeed “contact with the immediate Navarrese-Aragonese area”. As for the possible Italian influence over highly elaborated 15^th Castilian texts, we would like to point out an equally possible influence of Catalan. The 15^th c. is known as the “golden age” of Catalan literature, with its focus on Valencia, and it has been proved that some authors, specially Ausiàs March, had a great influence in later Castilian poetry. Furthermore, the 15^th coincides with the Compromise of Caspe (1412), which would entail a closer relationship between Aragon and Castile. In short, we are aware that proving this diatopic hypotheses is really difficult (maybe impossible) but we do not think it can be ruled out that easily.
5. Thank you for all your suggestions on English idioms and some linguistic passages. We consider that they have definitely improved our paper. We wanted to inform you that the paper had been already proofread by a native English speaker and a professional proofreader.

To conclude, we would like to repeat our thanks for the complete and interesting review.