Perfect-Perfective Variation across Spanish Dialects: A Parallel-Corpus Study

Round 1

Reviewer 1 Report

See the attached document.

Comments for author File: Comments.pdf

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

Review of “Perfect-Perfective variation across Spanish dialects: a parallel-corpus study”

Recommendation: Revise and Resubmit

The present manuscript uses a very interesting method (actually a combination of methods, which only becomes clear after reading it in detail) having a lot of potential for tackling intralingustic variation across the world’s Spanishes. However, I cannot recommend it for publication in its present form.

Before publication, the paper should be quite heavily reorganized. In its present form, several crucial issues are presented in a potentially misleading way.

1. The abstract wants to make us believe that the research consists of a corpus study of two dialects of Spanish. Actually, the research consists of a combination of two methods: corpus research and elicitation (probably with a questionnaire, but this remains implicit). And it is not two dialects of Spanish, but three regional written varieties of Spanish as reflected in very particular translations. One of them, Mexican, is later discarded, because it is found by confronting it with the result of elicitation that it is not representative of Mexican Spanish for the feature considered.
2. In the light of previous work on Spanish translations of Harry Potter (which is not discussed in the present manuscript), it cannot be taken for granted that the translations are representative of Castilian, Rioplatense and Mexican Spanish and the article actually verifies this (without admitting it explicitly) and comes to negative conclusions for Mexican Spanish. It must become clearer why the Mexican translation is excluded: because verification with native speakers reveals that the text is actually not representative for Mexican Spanish for the feature investigated (which raises the issue as to how certain we can be that the other two texts are representative, an issue that is not discussed).
3. Research on perfect within and across languages (except English) is underestimated (unnecessarily, since this issue is not crucial for the study). Typological work on perfects and other tense-aspect forms, to the extent mentioned, are adduced either for English or for methodological issues.
4. Given the focus on intralinguistic variability, it is strange that crucial parameters of intralinguistic variability such as modality (spoken vs. written), literary language (standard written language) vs. colloquial language etc. are not or not sufficiently addressed.
5. It does not really become clear what the method of multidimensional scaling really adds to more conventional methods such as frequency counts.
6. There must be a method discussion addressing the limitations of the methods used.

A major methodological problem of the manuscript is that it is taken for granted that translations of Harry Potter made in Madrid and Buenos Aires are representative of Castilian (European/Peninsular) and Rioplatense Spanish, respectively. In looking at features in the translations you can choose to

A: take the features as constant and consider the variety of the text as variable, or,

B: as in the manuscript, consider the varieties as a priori established and investigating the grammatical properties as variables.

C: recognize that both variety and features are variables

Fact is that most existing work, not all of which is Academic, opts for A, which suggests to me that the manuscript misleadingly takes for granted that the material considered IS representative for Castilian and Rioplatense Spanish

Consider, e.g.

Sundell, David. 2010. El español neutro en la traducción intralingüística. Un estudio sobre el uso del español neutro en las traducciones intralingüísticas de Harry Potter y la Orden del Fénix. Oslo: Master thesis. https://www.duo.uio.no/handle/10852/25810

see also

https://www.potterglot.net/harry-potter-and-the-spanish-tykes/

https://historiasecretadelmundomagico.com/

[I have no access to Tarantino, Patricio. Historia Secreta del Mundo Magico. It seems to be a popular rather than academic book, but it might contain a lot of relevant information, especially because it is written by an Argentinian writer.]

It seems to me that there might be a point in stating that the actual situation we have to deal with is C. It is not exactly known what kind of variability the translations reflect (what kind of Spanish) and it is not exactly clear how the features are distributed across varieties. This requires a triangulation of methods (more than one method). I consider it to be a strength of the manuscript that it actually uses multiple methods to tackle this difficult situation, but this is not reflected in the structure of the paper and in the way it is presented. To start with we have parallel texts where we can investigate a feature such as the perfect. However, we do not know whether the texts are representative of the varieties we think they might be representative of to start with. So, first, we have to check whether what we think might be Mexican Spanish etc. really is Mexican Spanish. The paper shows that this is unlikely for Mexican Spanish but likely for Castilian and Rioplatense. This methodological triangulation (multiple methods to tackle multiple unknown variables) must be turned into an explicit strength of the approach.

For earlier work on Spanish Harry Potter translations, see, especially, Sundell (2010: 51) “Como hemos mencionado, el uso del pretérito indefinido en lugar del pretérito perfecto predomina en el español de América. Consecuentemente, en la versión del Cono Sur es muy frecuente el cambio de los verbos en pretérito perfecto por verbos en pretérito indefinido, un cambio que no se hace frecuentemente en HA. Los cambios en CS son un tipo de cambio que corresponde con las características que presenta Bravo García (2008:45), acerca de usar el pretérito indefinido en lugar del pretérito perfecto; mas no con las observaciones de Petrella (1998:981-982), quién dice que predomina el uso del pretérito perfecto en los doblajes argentinos. Rodríguez Murguiondo (2010a) explica la decisión de cambiar los tiempos compuestos por tiempos simples:...”

and passim

Looking at this literature, it becomes clear that it is quite problematic to take for granted that the varieties considered represent such and such dialects. Also, the notion of “dialect” is rather vague; “dialect” can mean many things. But it seems to me that literary standard varieties in a pluricentric language is not the most prototypical use of the term. So, the notion of “dialect” needs at least some clarification. Sundell (2010: 8; 2) speaks of “español neutro en las dos traducciones intralingüísticas”. “Las traducciones que estudiamos en esta investigación son la versión del Cono Sur y la versión para el resto de Hispanoamérica.” In the light of this approach, it does not seem to me to be uncontroversial to speak of Spanish dialects here without any discussion.

It seems to me that the paper heavily underestimates the work done on intralinguistic variation of the perfect and related forms in languages other than English. I would like to mention here just two papers addressing Italian and Basque

Squartini M., & Bertinetto, P.M. 2000. Tense and aspect in the languages of Europe. The Simple and Compound Past in Romance Languages. In Dahl, Ö. ed., 2000. Tense and Aspect in the Languages of Europe, 403-439. Berlin: Mouton de Gruyter,

esp. Section 5. The chapter is also highly relevant for Spanish, of course.

Aldai, Gontzal. 2007. Discreteness and non-discreteness in the design of tense-aspect-mood. In New challenges in typology: Broadening the horizons and redefining the foundations: 271-291. Berlin: Mouton de Gruyter.

But there is certainly much more relevant work. The problem is not that this literature is not mentioned (well, Squartini & Bertinetto 2000 probably should), but that the paper wrongly suggests that such work does not exist.

Running Comments

p.1;9 and passim “Castilian Spanish” is ambiguous: can be used as the authors use it as the variety of Peninsular Spanish spoken in northern and central Spain, but also as the standard form of Spanish. This ambiguity must either be addressed or – better – another less ambiguous term should be used: Peninsular/European Spanish?

Please define “Rioplatense” somewhere. Not every reader will know what it is and especially what exactly is included.

p1. Examples for  Pretérito Perfecto  Compuesto PPC and Pretérito Indefinido PI should be given early in the text. Preferably in the form of some glossed examples.

2;50 This paper is thus structured as follows > This paper is structured as follows (Do not use thus where something does not strictly follow)

2;65 say explicitly somewhere what small caps in perfect, past stands for

2;75 The passage suggests that Comrie 1976 and Bybee, Perkins and Pagliuca 1994 are descriptive literature about the English present perfect, and this is entirely mistaken. Possibly those works address the English present perfect in some minor passage and then page numbers should be given.

2;78 I do not think Mary has read Der Zauberberg is a prototypical example of a result achieved. Very unclear to me what “so now she has that knowledge” means in connection to that novel.

2;91 “One problem with this approach has been that most of this research has focused on English and English alone.” Is it fair to adduce typological literature such as Comrie 1976 and Bybee, Perkins and Pagliuca 1994 and then blame them for only addressing English?

2;98 I do not think French can be addressed without elaborating on which forms French has to express the perfect.

3;104 “(Ana ha desayunado en ese café 104 lit. ‘Ana has eaten breakfast at that café’), some past actions are conveyed with a Perfective 105 Past form, the PI (Ana desayunó en ese café ‘Ana ate breakfast at that café’).” These or any examples should come earlier and they should be glossed with interlinear glosses.

2.2 The authors argue here that PI and PPC are understudied in Spanish. OK, everything is understudied, but what is really the evidence for this claim? What kind of literature search has been made by the authors? Has the literature in Spanish been surveyed?

I made a simple search for “perfect tense Spanish” in Google scholars. This search will miss all the literature in Spanish and any other languages but English. Here are some of the top results. Those marked with X are mentioned, but not the other ones.

Cowper, E., 2005. The geometry of interpretable features: Infl in English and Spanish. Language, pp.10-46

Escobar, A.M., 1997. Contrastive and innovative uses of the present perfect and the preterite in Spanish in contact with Quechua. Hispania, pp.859-870

X Howe, C. and Schwenter, S.A., 2003. Present perfect for preterite across Spanish dialects. University of Pennsylvania Working Papers in Linguistics, 9(2), p.7.

X Howe, L.C., 2006. Cross-dialectal features of the Spanish present perfect: A typological analysis of form and function (Doctoral dissertation, The Ohio State University).

Pérez, N.J., Non-standard configurations in Eonavian Spanish: how to be perfect without a perfect tense.

Said, S.E.S., 1976. VARIATION IN USAGE OF THE PRESENT PERFECT TENSE IN THE SPOKEN SPANISH OF MEXICO CITY. The University of Texas at Austin.

X Schaden, G., 2009. Present perfects compete. Linguistics and Philosophy, 32(2), pp.115-141.

X Westmoreland, M., 1988. The distribution and the use of the present perfect and the past perfect forms in American Spanish. Hispania, 71(2), pp.379-384.

Yupanqui, I.M.J., 2006. The use of the preterite and the present perfect in the Spanish of Lima (Doctoral dissertation, University of Pittsburgh).

[I am entirely leaving out the massive literature on L2]

This is not a problem per se, but it is a problem in connection with the claims

“In any case, most of this research has been done in Castilian dialects” (if Studies dealing with other varieties are not mentioned)

E.g. J. Yupanqui is referred to for “cross-dialectal variation”, but not for her research on Peruvian Spanish perfects.

3;122 “We de-122 scribe them directly below.”: describe is the wrong word here: survey or summarize.

3;128 Moreover, older work is primarily 128 based only on constructed examples,

This is a very general statement. Is it really true? As claims made elsewhere in the paper are overstated, I am at a loss as a reader here whether I really can believe what is claimed here without any evidence adduced.

3;132 “hodiernal definite past adverbials” what is this exactly? Please, define the term. Same for “pre-hodiernal”

“some regions in Spain (Canary Islands, Galicia,” Does “Galicia” here refer to Galician Spanish or to Galician?

3;143 However, the PPC is 143 said to appear in experiential indefinite past contexts, without a link-to-the-present com-144 ponent (Rodríguez Louro 2009, 2012). This author

Strange to refer back by “this author” to only a parenthesis.

3;150 A potential source of variation within this domain can be: what is meant by “this domain” here?

3;157 “seem to indicate different constraints at play across 157 the dialects, ultimately producing diverse diachronic trajectories” I do not understand what exactly is meant by “constraints” here.

3;157 “In summary, as we can distinctly see, the picture on the distribution of the PPC and 159 the PI across dialects of Spanish is far from clear.” Strange way of putting it.

3:160 “To bridge this gap between macro and 160 micro-variation,” It does not seem to me that it is clearly worded here what exactly is the gap and why the study of these two “dialects” are an appropriate solution.

3;165 “Dahl and Wächli 2016”: This is a typological study of the perfect. Why is it referred to for method only?

Section 3

I would expect here that modality is addressed here: spoken vs. written language.

If Southern Cone is equated with Rioplatense, this should be made explicit.

“The cities where the books were pub-173 lished indicate that these areas respectively correspond to the dialects of Madrid (Castil-174 ian), Mexico City (Mexican)1 and Buenos Aires (Rioplatense).”

Can you give us some more information about these publications? Actually, no references to the three books are listed anywhere. Sundin (2010), e.g. tells us that all books have the same publisher Salamandra:

“Sin embargo, la mera realidad es que la Editorial Salamandra, responsable de la distribución de Harry Potter en español, incluye a Argentina, Chile y Uruguay en el Cono Sur, mientras los otros países hispanoamericanos, incluido los EE.UU., reciben la otra versión adaptada (Congil 2010).”

But in https://www.potterglot.net/harry-potter-and-the-spanish-tykes/ we see that things are more complex. Is it really true that the Rioplatense version was published in Buenos Aires?

This has been an issue for collectors: “There are other reasons that I would think collectors would find it unique aside from having the “first edition” thing going for it. Given the history of the publishers that I discussed previously, having an Emecé imprint is definitely rarer than a Salamandra imprint. But also, this edition is the only one that carries an Argentinian ISBN number (978-950-04-1957-4); all the other Emecé imprints, despite carrying the name, were produced through their (at the time) Spanish subsidiary Salamandra and so all have Spanish ISBN numbers (ex. 978-84-9838-709-4). (Actually, there are also library / school editions farmed out to other publishers that inexplicably have the English language “0” identifier (ex. 978-0-320-03782-5) but that’s a bit of a tangent.)”

What we need is clear references to the books that were actually used for the investigation.

4;181 We extracted all cases where either the original in English uses a Present Perfect: in dialogue only?

4:185 “the Perfect Extractor, a 185 system developed by van der Klis, Le Bruyn and de Swart (2017), which searches the text” The text in which language? English or Spanish or both?

4:205 “Since inferring generalizations from a set of numerous contexts in four different 205 translations can be quite demanding, we also make use of a cartographic approach to data 206 visualization known as Multidimensional Scaling”: demanding for whom?

4;215 “a baseline map” what is a baseline map?

6;235 “of the distribution between Perfect and Past forms” This should probably be Past perfective. Generally, the terminology is not consistent across the paper. Why are the abbreviations PI and PPC not used here?

7;248 “retraction”: Retraction is a diachronic term. This is no diachronic study. So I do not understand the term retraction here, what is meant by the term?

similarly “remain” in 7;249

7;253 “we find that PPC use is greater in the 253 Mexican variety (n = 37), than in Castilian (n = 32)” “greater” too colloquial here and unclear what it means? “more instances”?

7;258 “the maps allow us to uncover a pattern across the different 258 dialects of Spanish:” It does not become clear what exactly the maps allow us to uncover. If the issue is frequency, this can be addressed easily by more conventional methods such as frequency counts represented in a table.

7;271 What I am missing here is a discussion of how the dimensions in the figure should be interpreted. What is Dimension 1 and what is Dimension 2? It may, of course, be possible that this cannot be said, but even then it would be useful to have some discussion of this issue.

7;274 “Mexican Spanish has a more extended use of the PPC than Castilian 274 Spanish, and this latter dialect, in turn, displays a broader use of this form than its Rioplat-275 ense counterpart.”

These statements are about particular texts and it has not been shown that these texts are representative for Mexican Spanish etc. The conclusions should be formulated more cautiously.

It also seems to me that it is wrong to say that MDS shows differences in frequency. MDS is not a frequency count method.

7;280 “where native speakers had to choose be-281 tween the PPC and the PI forms in a context-sentence pair.” This change of method and data comes unexpected and it is not clear to me whether “interim discussion” is the right section for it.

If data collection was done with a questionnaire, it would be good to add the questionnaire as an appendix.

8;286 “We created 2 lists of 15 items each, and asked 286 native-speakers of each dialect (n = 10 per dialect)”

How were the speakers selected? How is it assured they are representative for the dialect? What kind of instructions did they receive? We know from variationist/sociolinguistic research of the Labovian type that each speaker has access to various styles and the elicitation design can have strong effects on which forms are produced.

8;291 Table 1

I do not understand the change of method from MDS to standard frequency tables. The elicited examples seem to be places in Harry Potter. Why isn’t their location shown on the semantic map produced by MDS?

8;291 Table 1

I don’t understand the table. What is shown by columns, what is shown by rows?

8;295

“the corresponding dialect-specific translation”: it cannot be taken for granted that these are dialect-specific translations.

8;295

It would be interesting to learn more about the variability between the participants. The data collected seems to allow for investigating what kind of variation there is between participants of a variety as opposed to between varieties.

8;303

“We consider that these results constitute a firm basis for leaving the Mexican Spanish 303 translation out of our subsequent analysis”

This is an extremely strange way to put it. It seems to me that the results show that the assumption that a Harry Potter translation made in Mexico is representative of a Mexican dialect of Spanish was mistaken. And this raises doubts also about the translations made in Madrid and Buenos Aires.

8;308 “subset relation” I do not understand what is meant by subset here and it does not become clear to me from the following discussions in which ways we have to deal with subsets here.

• 7 I do not really understand why each variable is considered one-by-one, rather than for instance one heat-map analysis for all variables together.

13;480 “Results from this coding indicate that in relation to inherent aspect, durative contexts 480 increase the chance that Rioplatense Spanish displays a PPC instead of a PI, since both 481 dialects are not significantly different from each other in this kind of contexts.” I do not understand this statement. “since...” looks as a causal clause, but I cannot see any causal relationships between the clauses.

13;484 “However, 482 Castilian Spanish uses the PPC in all durative aspect cases, while Rioplatense Spanish only 483 uses it in a bit more than 50%. It” > the Castilian Spanish translation > the Rioplatense Spanish translation

It does not become clear to me here why the features are considered one by one rather than in a combined integrated model. The different variables are probably not independent of each other and by combining them in a common model it might be possible to assess whether any relationships are spurious. See, for instance:

Levshina, N. 2016. Why we need a token-based typology: A case study of analytic and lexical causatives in fifteen European languages. Folia Linguistica 50(2). 507–542.

13:507 “maintain” is the wrong word here. It suggests a diachronic perspective that you cannot provide.

The section “Discussion and general conclusion” is partly too much on a result level for being a general conclusions. Such statements as “Castilian Spanish uses a PCC in 10 cases while Rioplatense Spanish 506 only maintains that marker in one case” belong into a result section and not into a conclusion section. The paper would profit from a more succinct conclusion section which is nothing else but conclusion (possibly with cross-references to earlier reported results).

There is a lack of discussion what as few as 10 examples etc. in one particular text really mean for Rioplatense and Castilian Spanish. Put differently, there is no critical method discussion.

13;525

“As for the hodiernality constraint, it is also important to look at previous research 525 (e.g., Harris 1982) that indicates that”: shouldn’t this have been introduced much earlier in the paper? Conclusions do not usually contain new background information.

The paper might profit from one or several appendices. Consider e.g. the Appedix in Sundell (2010).

Comments for author File: Comments.pdf

Author Response

Please see attachment

Author Response File: Author Response.pdf

Reviewer 3 Report

El trabajo que se presenta para evaluación constituye una relevante aportación a la dialectología del español. Pese a los numerosos trabajos publicados sobre el tema, todavía no se han aclarado suficientemente las diferencias que determinan el empleo de los tiempos compuestos en las diferentes regiones dialectales hispánicas. Esta laguna bibliográfica es especialmente importante en lo relativo al uso de los tiempos compuestos en América, donde las diferencias entre países (y regiones dialectales) y respecto del español de España son muy marcadas. Por lo tanto, la aportación que se somete a evaluación tiene una alta relevancia.

El trabajo tiene una buena fundamentación metodológica, tanto en lo relativo a la extracción de ejemplos como en lo referente a la anotación y al análisis. Con todo, me asalta la duda de si los gráficos que el autor incluye en su trabajo pueden resultar de difícil comprensión para alguien no habituado a este tipo de representación de resultados cuantitativos. A fin de ayudar a la comprensión de los datos, recomiendo una explicación algo más detallada.

Para contextualizar adecuadamente la alternancia de tiempos entre variedades dialectales, habría sido relevante incluir una pequeña revisión bibliográfica sobre la historia de los tiempos compuestos en América. Como Octavio de Toledo (2021) ha probado, no existía diferencia entre América y España hasta después de los sucesivos movimientos de independencia de los países americanos. Posiblemente, es el momento en el que el español empieza a ser adquirido por un número importante de hablantes nativos de lenguas amerindias cuando se empiezan a gestar las diferencias. Para el empleo sincrónico de los tiempos compuestos en América es relevante leer la aportación de Bermúdez (2005) sobre el valor evidencial que este autor atribuye a los tiempos de pasado en Argentina. También contiene explicaciones interesantes el trabajo de Soto (2015). En lo referente a la revisión bibliográfica sobre la distribución de tiempos en la Península, faltan algunas referencias básicas, todas ellas más recientes que la que se cita en el trabajo (Serrano 1995). Recomiendo la lectura de los trabajos más recientes, pues creo que permitirá al autor añadir a su trabajo datos relevantes para su trabajo. Incluyo las referencias en el apartado bibliográfico final.

Aspectos formales

El autor debe prestar atención a algunas leyendas de gráficos que aparecen en español y no en inglés.

Referencias bibliográficas

Azpiazu Torres, Susana (2019): La composicionalidad temporal del perfecto compuesto en español: estudio sincrónico y dialectal, Boston / Berlín, De Gruyter.

Bermúdez, F. (2005). Los tiempos verbales como marcadores evidenciales: El caso del pretérito perfecto compuesto. Estudios Filológicos. https://doi.org/10.4067/s0071-17132005000100012

Camus Bergareche, Bruno (2017): “Perfecto simple y perfecto compuesto en los
atlas lingüísticos españoles”, Moenia, 23, págs. 285-316.

Llorente Pinto, María del Rosario (2017): “Los pasados perfectos en Ávila
y Zamora (según las encuestas realizadas para el ALEP)”, Orillas, 6, pags.
441-453.

Octavio de Toledo, Álvaro, (2021): “Avances recientes en la investigación diacrónica del pretérito perfecto compuesto: una visión panorámica y cuatro observaciones variacionales”, en Quijada van den Berghe, Carmen; Gómez Asencio, José J., Los pretéritos perfectos simple y compuesto en español peninsular y en otras lenguas románicas. Madrid: Arco Libros.

Soto, G. (2015). El pretérito perfecto compuesto en el español estándar de nueve capitales americanas: frecuencia, subjetivización y deriva aorística. In S. Azpiazu (Ed.), Formas simples y compuestas de pasado en el verbo español (pp. 131–146). Editorial Axac.

English version has attached. (translated by google)

Comments for author File: Comments.pdf

Author Response

Please see attachment

Author Response File: Author Response.pdf