Review of The Morphosyntax of Portuguese and Spanish in Latin America , by Mary A . Kato

This volume, part of the series Oxford Studies in Comparative Syntax, contains fourteen chapters, in addition to a foreword by the editors. It is devoted to the formal analysis of different aspects of American varieties of Portuguese and Spanish, while presenting their innovations with respect to their European counterparts. The papers collected have been presented at a meeting of the Romania Nova project and represent a wide range of topics in syntax, including null subjects and objects, complex predicates, clitics, and operator-variable constructions, besides one paper on morphology (truncated nominals). They explore these themes according to different theoretical assumptions and methodological tools.


Introduction
This volume grew from papers presented at a meeting of the Romania Nova project, organized by the same editors of the book.Although the project is affiliated with the generative enterprise, and more specifically with the Principles and Parameters framework, it also accepted contributions associating methodological tools from sociolinguistics and historical linguistics with formal assumptions, an eclectic point of view that is very welcome in face of the discussion of innovative varieties.Anyway, the interested reader should ideally have an advanced degree of knowledge in formal grammar, and especially in Syntactic Minimalism, in order to fully grasp the contents of this book.
This volume represents a very welcome addition to the area of comparative syntax, and more specifically to the subfield termed Microparametric Syntax, dedicated to the study of dialects or close grammatical systems (cf.Kayne 1996).First, because Spanish and Portuguese come to be languages so close that some phenomena may be analyzed either intercontinentally or regionally, in the areas where they are spoken.Second, because Spanish and Portuguese dialects have different characteristics in the Latin American context: (i) their sociolinguistic diversity is heavily based on diastratic variation; (ii) many of their particularities came out from situations of "deficient" second-language acquisition, possibly more than from a variety of "imported dialects" from Europe (cf.Holm 2004: 15 -19, among others).This shows that the phenomena observed in Latin American dialects tend to be very different from those discussed in studies about European dialects (cf.D'Alessandro, Ledgeway & Roberts 2010: 1-51, among others).Of course, the study of French dialects in North America may be incorporated in such a dialogue as well, as noticed in the foreword.
The editors have recognized a lack of systematicity on how dialects are referred to in the book, which can be partly attributed to the lack of efforts toward studies on syntactic dialectology, besides the fact that traditional dialectology is still an area in need of further development in the continent.Nevertheless, the generative approach of using intuition data alone, followed in most of the papers, may lead some readers to have doubts not only regarding the range the observations made apply to, but sometimes also regarding their very exactitude, since they may be subject to syntactic priming (cf. Snyder 2000).This is especially important in areas where the effects of language contact and/or schooling might influence judgments.Thus, it would be very enriching if the contributors followed a concerted strategy on how to cope with dialectal data.

The chapters
The book has fourteen chapters, in addition to a foreword by the editors.The chapters follow a thematic order, which is explained in the foreword: subjects (chapters 1 to 5), complex predicates (chapters 6 and 7), objects (chapters 8 to 10), wh-constructions (chapters 11 to 13) and truncated nominals (chapter 14).
In chapter 1 ("Microparametric variation in Spanish and Portuguese: The null subject parameter and the role of the verb inflectional paradigm"), Maria Eugênia Duarte and Humberto Soares da Silva confirm the hypothesis according to which null subjects correlate with strong inflectional paradigms, although they question that this can be explained through functional compensation.Using the methodological tools provided by quantitative sociolinguistics, the authors observe the following hierarchy among the studied languages: a prototypical null subject language (Italian), "not so null subject languages" (Peninsular Spanish, European Portuguese, Argentinean Spanish), and languages with grammars showing change in the direction of non-null subject languages (Brazilian Portuguese and Puerto Rican Spanish).The categorizations are based on the factor groups selected as significant for each variety/grammar.
In chapter 2 ("The null subject parameter revisited: the evolution from null subject Spanish and Portuguese to Dominican Spanish and Brazilian Portuguese"), José Camacho aims at reframing the classical requirements for the licensing of null categories, interpretability and identification (cf.Rizzi 1986) in the minimalist framework, in terms of the interpretability of formal features in inflection (I) and the existence of a topic antecedent to the null subject, respectively.Therefore, he develops a variant of the pronominal inflection analysis, while adopting Pesetsky & Torrego's (2001) model together with an antilocality constraint according to which there cannot be an agreement relation between a specifier and its own head.A natural consequence of this system is that either overt subjects cannot appear in Spec,IP, i.e. they become dislocated, or pronouns become unvalued (expletives).Each of these strategies represents solutions adopted in Brazilian Portuguese and Dominican Spanish, to give just an example of a consequence of the analysis.
In chapter 3, "On the notion of partial (non-) pro-drop in Romance", Andrés Saab explores the idea that null subjects in null subject languages (NSLs) are instances of morphological ellipsis, by comparing them with Northern Italian dialects (NIDs) and with Brazilian Portuguese (BP), which express different types of partial pro-drop.He adopts the deletion approach to null subjects proposed in Saab (2008).The author contends with the idea that agreement is interpretable in null subject languages in view of the lack of strict identity effects under ellipsis.The distinction between the two former systems is developed under the framework of Distributed Morphology: clitic subjects in Northern Italian dialects cannot be elided due to a ban on independent subword deletion.The latter system is a consequence of losing the syntactic specification for person features in I, ellipsis being resolved in the syntax.
In chapter 4 ("Rebel without a Case: Quantifier floating in Brazilian Portuguese and Spanish"), Renato Lacerda explores the following question: why is it that only quantifier counterparts of 'all' and 'each one' can float in Romance languages?His proposal is that QP-type quantifiers do not have to check Case in the narrow syntax.By doing so, he asserts that differences in the informational status of quantifier and DP are not the only factors determining the occurrence of quantifier floating, pace Valmala (2008).In line with contextual approaches to phasehood, he assumes that only the highest DP projection assumes phasal status, and explains restrictions to quantifier floating at the immediate postverbal position (a low focus position).By conjoining two syntactic principles (antilocality and phase impenetrability), Lacerda also explains the distribution of tudo, a "colloquial" counterpart of todos without agreement.For him, this form shows a stress requirement linked to focus.He finally explores some differences in the distribution of quantifier floating in BP and Spanish.
Chapter 5, "Subject and topic hyper-raising in Brazilian Portuguese: a case-study on reference sets for economy computations" by Jairo Nunes, investigates the problem created to economy computations by a combination of different properties of BP: double subjects, possessor/locative raising to Spec,IP, subject hyper-raising and topic hyperraising: in view of the last resort requirement, one would expect that some derivations should be impossible, against the facts.The proposed solution is that derivational economy is computed in a local manner i.e. from different distributions of the lexical items in the numeration.
In chapter 6 ("<V-and-V> constructions in Portuguese: the case of <ir-and-V>"), Madalena Colaço and Anabela Gonçalves focus on the analysis of the <ir-e-V> construction, a complex predicate formed by 'go' and 'and' plus a lexical verb in two varieties of Portuguese.They show that the construction makes a contribution to the narrative, where ir is similar to a raising verb, and e is taken to be the lexicalization of tense (T) on the complementizer head (C).They analyze the construction as a subcase of hyper-raising.The differences between BP and EP relate to the greater variability found in the latter.
In chapter 7 ("Inflected infinitives and restructuring in Brazilian Portuguese"), Marcello Modesto aims to show that Landau's (2000) partial control versus exhaustive control classes correlate in BP with the possibility of inflected infinitives being restricted to the latter class of verbs.This mimics the structural differences between CP versus vP (possibly also involving other projections such as AspP, but crucially not IP) and the availability of phenomena relating to "restructuring" versus "non-restructuring", respectively, in accordance with Wurmbrand (2001) and Grano (2012).
Chapter 8, "The null object in Romania Nova" by Sonia Cyrino, advances an analysis for null objects in BP and shows that this phenomenon is distinct from the one found in varieties of South American Spanish.For her, BP null objects would be the result of DP ellipsis because [-animate] referents and sloppy readings are available.This is explained in terms of Platzack's (2003) nanosyntactic theory: null objects in BP are complements of RootP, holding a Path or Theme role.Accordingly, the verb moves to a lower position, from where it can license ellipsis.In South American Spanish varieties, either null objects are the result of VP ellipsis affecting [+animate] referents -a phenomenon also available in BP -or they are subsumed under the more general cases of null objects observed in Cummins & Roberge (2005) for French.
In chapter 9 ("The external possessor construction in European Portuguese and Brazilian Portuguese"), Maria Aparecida Torres Morais and Heloisa Maria Lima-Salles examine external possession, which is nowadays restricted in spoken BP to the cases where the possessor is in subject position.The authors assume the theory of datives as expressions of applicative projections (Pylkännen 2002) and a constructivist approach to argument structure.They observe that external possessors in European Portuguese may express three different types of low applicative heads, expressing the Source, the Recipient, or the Locative role.The loss of external possession in BP is described as a consequence of the reanalysis of its pronominal system, together with the loss of the functional preposition a (replaced by the lexical preposition para in ditransitive constructions), illustrated with diachronic data collected from previous studies.
In chapter 10 ("Agreement and DOM with impersonal se: a comparative study of Mexican and Peninsular Spanish"), Francisco Ordóñez and Esthela Treviño discuss agreement properties of the patient/theme argument in impersonal constructions with se in transitive clauses.To do so, they observe the interaction of this construction with Differential Object Marking (DOM), together with its pronominalization properties.By comparing Mexican Spanish, Peninsular Spanish and other South American varieties of Spanish, it is shown that they all share the inexistence of a clitic counterpart for non-DOM objects in clitic left-dislocated constructions, whereas the form of the clitics varies between dative and accusative, depending on the dialect.Their analysis states that inherent Case is assigned to DOM objects, whereas nominative is available for non-DOM objects, a type of passive, the difference between true impersonal se and passive se having to do with the projection of a lower vP phase only in the former.
Chapter 11, "Some syntactic and pragmatic aspects of WH-in-situ in Brazilian Portuguese" by Maria Cristina Figueiredo Silva and Elaine Grolla, explores cases of obligatory wh-in-situ in BP, a language where this construction is normally optional and shows a broader distribution than in other languages.The contribution discusses proposals for the pragmatic licensing of the construction, showing that linguistic common ground disfavors wh-in-situ, on the basis of data from language acquisition experiments and sociolinguistic interviews.From a syntactico-semantic viewpoint, they discuss Kato's (2004) proposal, for whom ordinary in-situ questions result in fact from movement to FocP in the lower clause periphery.They also show evidence against this proposal, based on word order facts, concluding with a derivational problem related to the connection between a question morpheme and the wh-word.
In chapter 12 ("Pseudo-clefts and semi-clefts: an analysis based on Portuguese"), Mary A. Kato and Carlos Mioto claim that pseudo-clefts (PCs) and semi-clefts (SCs) are derivationally independent and are both interpretively monoclausal.The authors present old and new arguments showing that SCs cannot be derived from PCs.They present a new cartographic approach for the latter, considering that remnant movement of the lower VP is responsible for the focalized part of the clause, and IP remnant movement moves material into the presupposed part.
Chapter 13 ("From [o [que]] to [o que] in Brazilian Portuguese Free Relatives: a diachronic view"), by Paulo Medeiros Junior, explores free relatives in diachronic BP between the sixteenth and the nineteenth century.The author's main hypothesis consists of there being an evolution from semi-free relatives (where que is mapped to Spec,CP) into free relatives (where o que is in Spec,DP).This change process is described as a consequence of a cue-based reanalysis, possibly a subcase of grammaticalization.Medeiros Junior shows that the non-ambiguous context for semi-free relatives is very rare, being limited to those clauses including a preposition between o and the wh-word.He finally suggests that Spanish has followed a path of change very similar to the one shown for BP.
Finally, Ana Paula Scher offers in chapter 14 ("A study of truncated nominal forms in Brazilian Portuguese: their derivation and their relation to nonverbal form classes") a treatment for truncate nouns and adjectives in BP within the framework of Distributed Morphology.In this text, the author recovers the idea that nonverbal categories are also distributed according to morphological classes.She then compares truncated forms ending in -a with those ending in -as or -(i)s, holding that the special meaning of truncated nominals results from an evaluative morpheme which may be expressed by a real suffix.Some forms may exclude parts of the root due to a reanalysis of those sequences as suffixes by speakers, as -ej-in cerva < cerveja ('beer').

Evaluation
This book presents valuable studies that pave the way for a deeper understanding of some key morphosyntactic aspects of Portuguese and Spanish in the Latin American setting.Studies on Portuguese, and more specifically on BP, are more present than those on Spanish, although most of the chapters show a clear effort to compare Portuguese and Spanish varieties somehow.
As already mentioned in the introduction, the lack of some common methodology on data collection and on how to refer to the different dialects is an issue for most of the chapters.An important exception to this trend is chapter 10, where the authors present a brief note on data collection in the first paragraph, specifying the cities from which they took their data.This may have relevant consequences.For example, regarding the diatopic domain, many of the truncated forms cited in chapter 14 are common in Southeastern Brazil, and more specifically in the São Paulo state only (e.g.padoca ['bakery'] and bermas ['bermuda']).The reference to the diastratic variation is unsystematic: "colloquial BP", "popular BP" or "spoken BP" may refer to different groups of people, a fact that is not totally clear.This may be the source of disagreements in data judgments.For instance, as a native speaker of Brazilian Portuguese, I judge as unacceptable sentence (30c) from p. 90 (chapter 4), where tudo ('all') occupies the prenominal position: (1) (*) Eu comprei TUdo os livros.I bought everything the books 'I bought all the books.' The same goes for the sentences in (6) in chapter 5 (p.108), presented as examples of the interaction of topic hyper-raising with possessor raising to Spec,IP: these books seem.3pl that the author is unknown 'The author of these books seems to be unknown.' In both cases, the original judgments may be valid for other speakers, but since there is no note on data collection procedures, the reader infers that they reflect each author's intuitions.This is problematic because the use of tudo is highly variable in BP, as recent work has shown (Trannin 2016).Although hyper-raising and possessor raising are common to different social strata, the sort of data in (2) may reflect a case of syntactic priming.
In fact, this connects with another issue, the dialogue with previous texts, which could be more explored.In chapter 5, the work by Andrade & Galves (2014) is cited, but the difference stated there between "real subject topics" (in Spec,IP) and "spurious subject topics" (i.e.hanging topics) is overlooked.If it were taken into account (together with the alternative judgments), the need of explaining competing derivations in this area would be immediately eliminated, in view of the fact that hanging topics necessarily block verb agreement.The same goes for the lack of agreement in (45b) and (46b) in p. 122, presented in (3), which should be striking in view of the original judgments for (2) above: (3) a. *Os relógios quebraram o ponteiro deles.the watches broke.3pl the arm of-them 'The arms of the watches broke.' b. *Essas gavetas cabem muita coisa nelas.
these drawers fit.3pl many thing in-them 'Many things can fit in these drawers.' In chapter 7, Modesto's discussion of complex predicates also overlooks some important work made in Brazil on auxiliary verbs and clitic climbing, such as Cyrino (2010), Rech (2011) and Andrade (2010).For instance, the application of Wurmbrand's (2001) ideas to "restructuring" contexts to diachronic European Portuguese can be found in the latter, and specifically a similar test on Negative Polarity Items on p. 168, which leads to the conclusion that there is constituent negation in the infinitival complement.
Other general methodological problems may be pointed out.For instance, in chapter 1, no results on social stratification are taken into account as an auxiliary tool to ascertain a change in progress regarding null subjects.In chapter 13, the author does not provide the list of primary sources used in his diachronic study, whence one cannot verify if the corpus includes language effectively "produced in Brazil".On top of that, some examples written by Portuguese authors that have never set foot on Brazil (p.316) make the reader believe that the development is not specific to BP.Other minor problems relate to the lack of careful copyediting, thus leading to some remaining typos and unnatural English expressions.Some chapters cite works that do not appear in their reference lists, such as Duarte (2000) in chapter 2, and Marchesan (2008) in chapter 13.
By pointing out these minor flaws and unanswered questions, I do not intend to detract the relevance of the collected works, but to show that their impact could be even bigger should more attention to methodological principles be paid.Some contributions present indeed nice consequences for parametric theory and beyond.For instance, chapter 1 confirms the fact that Italian is different from other Romance languages that are indisputably categorized as "consistent null subject languages", such as European Portuguese.In chapter 8, the reinterpretation in formal terms of a section of the Referentiality Hierarchy (a variant of functional hierarchies named in various ways since Fillmore 1968) makes precise statements about the typology of null objects.The reflections on wh-in-situ in chapter 11, although not translated into a specific proposal, involve a very useful discussion about the need of bringing different sorts of evidence into syntactic theorizing.
All in all, the book is very rich in analyses and well-balanced regarding the coverage of different topics in syntax (and morphology, to a lesser extent), providing a nice introduction to current research on Portuguese and Spanish in the Latin American context.Therefore, I strongly recommend it to researchers interested in Romance grammar, working on any theoretical persuasion.