The study also highlights the constructs of current linguistic theory, arguing for distinctive features and the notion 'onset' and against some of the claims of Optimality Theory and Usage-based accounts.
The importance of Henk Zeevat's new monograph cannot be overstated. [...] I recommend it to anyone who combines interests in language, logic, and computation [...]. David Beaver, University of Texas at Austin
Date: Sat, 28 Jan 2006 13:12:43 +0100 From: Joaquín Romero <firstname.lastname@example.org> Subject: Laboratory Approaches to Spanish Phonology
EDITOR: Face, Timothy L. TITLE: Laboratory Approaches to Spanish Phonology SERIES: Phonology & Phonetics PUBLISHER: Mouton de Gruyter YEAR: 2004
Joaquín Romero, Department of Anglogermanic Philology, Universitat Rovira i Virgili, Tarragona, Spain.
This book presents a selection of papers from the Laboratory Approaches to Spanish Phonology conference, held at the University of Minnesota in September 2002. The collection of papers is edited by the conference organizer, Timothy L. Face, who is also the author of one of the papers. The volume follows the philosophy of the conference in that it is an attempt to present a sample of current research in the phonetics/phonology interface of Spanish and it follows the example of the now consolidated Laboratory Phonology series of conferences and edited volumes. Following a brief introduction by the editor, the individual papers are presented in three major parts. Part 1 deals with Intonation, Part 2 is entitled Syllables and Stress, and Part 3 is presented under the heading Segmental Constraints. Following is a summary of each of the papers, with a brief critical evaluation stressing the major contributions as well as any possible shortcomings.
SYNOPSIS AND EVALUATION
Conxita Lleó, Martin Rakow & Margaret Kehoe, Acquisition of language-specific pitch accent in Spanish and German monolingual and bilingual children. The paper by Lleó, Rakow and Kehoe investigates the acquisition of some basic intonational patterns by monolingual and bilingual children. A comparison is made between monolingual Spanish and German children, on the one hand, and Spanish/German bilinguals, on the other hand, in terms of their acquisition of the basic patterns of prenuclear accents in broad-focus declarative utterances. The results show that, while the monolingual children seem to show no difficulty in acquiring the correct intonational pattern at an early age, the acquisition of these patterns in the bilingual children is more subject- dependent. One of the bilingual children seemed able to differentiate the patterns of the two languages, while the other used a variety of patterns in either language. These findings are taken as evidence that the acquisition of intonational patterns follows along similar paths to other prosodic features. They also give indirect evidence that unmarked patterns of intonation (H*L) seem to develop earlier than marked patterns (L*H). A drawback of the study is the very limited number of subjects that participated in it, which seems more suitable for a pilot study than a full-fledged one, especially since no mention is made in the paper of any possible further experiments. All in all, however, the paper provides interesting data of an aspect of L1 acquisition that is still largely unknown, especially as concerns bilinguals, and it opens numerous possibilities for further research in this topic.
Pilar Prieto, The search for phonological targets in tonal space: H1 scaling and alignment in five sentence-types in Peninsular Spanish. Pilar Prieto's contribution focuses on the description of sentence initial peaks as a function of sentence type. She compares the tonal scaling and alignment of five types of sentences in peninsular Spanish: statements, yes-no questions, wh-questions, imperatives and exclamatives. The main theoretical assumption in this paper is that variations in pitch range are not exclusively paralinguistic, they can also convey linguistic information, in particular, information about sentence type. Results confirm this hypothesis, showing that the scaling of initial tones is consistently higher in questions (both absolute and wh) than in statements. Similarly, imperatives and exclamatives show consistently higher initial tones than the other sentence types. In terms of tonal alignment, the results show a clear distinction between late H1 peaks for statements and questions, and an early H1 peak for imperatives and exclamatives. The author argues that these results evidence some shortcomings in the traditional autosegmental description of pitch accents. She proposes that certain extra features are needed to refine both tonal scaling and alignment, such as [delayed peak] and [raised peak]. While this proposal might be in agreement with current work in the phonology of intonation, it seems unnecessarily abstract and categorical. A more phonetically based approach might consider the possibility of looking at tonal scaling as gradient, so pitch accents need not be categorically described by specific features, but rather determined dynamically in relation to the overall tonal structure of the utterance.
Erik W. Willis, Dominican Spanish absolute interrogatives in broad focus. In his paper Erik Willis sets out to show that the intonational contours of broad-focus absolute interrogatives in Dominican Spanish differ from previously reported Caribbean Spanish patterns. Standing accounts of the intonation contour of these types of sentences in Caribbean dialects mention a pattern with an initial rise which is followed by a high plateau and ends in a fall. Instead, Willis provides experimental evidence that Dominican Spanish absolute interrogatives show an overall rising contour, with a characteristic upstepped prenuclear low pitch accent and a final plateau. These results are obtained from both laboratory speech and a spontaneous speech sample, which corroborates the validity of the findings as a faithful representation of Dominican Spanish intonation. Despite the rather narrow scope of this study, the author makes an interesting observation regarding the pragmatics of specific intonational contours, which suggests that the same pattern can be used in different dialects for different communicative purposes. In addition to possible implications for crossdialectal intelligibility, this point raises interesting issues having to do with the potential phonological nature of intonation contours and the existence of intonational universals.
David Eddington, A computational approach to resolving certain issues in Spanish stress placement. In his paper, David Eddington advocates for an analogically-driven or exemplar-based approach to phonological processing. Eddington investigates the relevance of this theory in reference to the aspects that determine the position of word stress in Spanish. Using materials from a Spanish frequency dictionary, he runs a series of computational simulations with two different analogical learning modeling techniques (Tilburgh Memory Based Learner and Analogical Modeling of Language). The results show that phonological notions such as syllable weight and the CV tier do not play an important role in the assignment of word stress in Spanish. Instead, what seems to determine whether word stress in Spanish will be antepenultimate, penultimate, or final, is the phonemic makeup of the word. Similar results are obtained from a different experiment in which native speakers of Spanish were asked to assign stress to a series of written nonce words. These results evidence the need to move away from models of phonological processing that are excessively mentalistic and abstract. At the same time, however, one should be cautious in drawing conclusions based on simulation data and nonce word data. For example, it is rather striking that, as illustrated in Table 2, two nonce words with identical phonemic makeup—'dagola' and 'dalona'— obtain opposite values for percentage of antepenultimate vs. penultimate stress. In a case like this, more detail about exactly how these tests were administered (i.e., the instructions that the participants received, the time allowed for the test, how the words were presented, etc.) is crucial in order to validate the findings unequivocally.
Timothy L. Face, Perceiving what isn't there: Non-acoustic cues for perceiving Spanish stress. The second paper in Part II is by Timothy L. Face, the editor of the volume, and it deals with the role of non-acoustic factors in the perception of stress in Spanish. The paper presents the results of a series of experiments that test whether factors such as syllable weight, lexical similarity, lexical regularity and morphological category play a role in determining stress placement. The experimental methodology is based on the use of synthesized trisyllabic nonsense words in which the acoustic cues that are known to correlate with stress are kept constant, i.e., pitch, duration and intensity; listeners are asked to indicate which syllable bears the stress. Contrary to the results found in Face (2000), the current study finds little evidence to support the role of syllable weight in the perception of Spanish stress. Instead, it is argued that it is the presence vs. absence of a word-final consonant that determines whether the word will be perceived as having final vs. penultimate stress. Other results point at the important role that the lexicon plays in the perception of Spanish stress. All in all, the results of Face's study are taken as evidence that stress placement in Spanish relies on many different factors other than just acoustic cues. In particular, a great deal of weight is given to the role of lexical patterns existing in the speaker's mind that influence the choice of stress pattern in nonsense words. One possible criticism has to do with the rather sparse presentation of the results as simple overall percentages, with little or no explanation of any possible within- group variation. In particular, as refers to the experiment dealing with lexical similarity, it is surprising that all the words in the list should behave the same way. One of the words in this list, 'capita', which is supposedly unequivocally reminiscent of 'capital' or 'capitán', can actually be a real word, as in the common expression 'per capita'. It would have been interesting to see whether this particular item behaved the same way as the rest.
Sharon Gerlach, Effects of environment on L2 epenthesis: Evidence for transfer of ranked constraints. Vowel epenthesis in Spanish is the subject of Sharon Gerlach's contribution to this volume. In this study the author combines experimental data with a theoretical account within the optimality framework. The experimental study looks for patterns of epenthetic /e/ in L2-English productions of native Spanish speakers in word- initial /s/+stop sequences. The dependent variable in the study is the preceding context, identified as 'vowel', 'consonant' or 'pause'. The results show some interesting patterns with respect to the dependent variable, but also, for example, in the separation of the subjects between 'epenthesizers' and 'non-epenthesizers'. Unfortunately, the paper does not investigate what factors might underlie this distinction. It does explore in detail, however, the effect of context and it concludes that, contrary to prior belief, a preceding consonant does not necessarily favor epenthesis over a pause. Preceding vowels, on the other hand, seem to hinder epenthesis, which is in accordance with previous research. The paper then moves on to providing an optimality-theoretic account of the phenomenon taking as a basis the interaction between resyllabification principles and the constraints governing the shape of onsets in Spanish. Independently of the validity of this analysis, it presents a simplistic view of the reality of the phenomenon by reducing it to an all-or-nothing situation. This is supported by the fact that no explanation is provided as to the criteria used to decide whether epenthesis was present or not in the experimental data. A more thorough analysis would likely reveal a gradient phenomenon, which would probably not make for such an elegant theoretical analysis, but would definitely be closer to reality.
Mark Waltermire, The effect of syllable weight on the determination of spoken stress in Spanish. Mark Waltermire's paper uses some of the data in Face (2000) to further investigate the role of syllable weight in the assignment of stress in Spanish. The study consists in the analysis of the results of a series of written questionnaires aimed at obtaining a glimpse into native Spanish speakers' intuitions as to the placement of stress in disyllabic and trisyllabic words. The data are subjected to a probability analysis that explores the effect that the presence or absence of final consonants will have in determining final vs. penultimate vs. antepenultimate stress. The results overwhelmingly confirm Face (2000)'s findings that the presence of a coda consonant in the final syllable favors final stress, whereas a light (codaless) final syllable induces penultimate stress. They also show, however, that the interaction between syllable weight and stress placement is more complex than that. Thus, his findings point at an intricate pattern of dependency between syllable weight and stress that can hardly be handled by traditional rule-based approaches. It is surprising, however, to see such a strong claim for the role of syllable weight in the same volume where Face admits that his own claim in this sense in Face (2000) was misled by a flaw in the experimental design.
Travis G. Bradley, Gestural timing and rhotic variation in Spanish codas. In this paper, Travis Bradley deals with the phonetic variation of rhotics in preconsonantal position in Highland Ecuadorian Spanish. He provides acoustic data that show how homorganicity with the following consonant results in an assibilated pronunciation of -r-, while when the following consonant is not coronal, the -r- is realized as the expected flap followed by a so-called svarabhakti vowel. This variation is explained within the theory of Articulatory Phonology in terms of gestural alignment. In the homorganic situation, the supralaryngeal gestures for -r- and for the following consonant are produced with the same articulator, which results in a process of gestural blending. In the heterorganic configuration, the -r- and the following consonants do not share a supralaryngeal gesture and therefore are pronounced in sequence. The temporal coordination between the two is responsible for the svarabhakti vowel. This simple explanation provides a lot more insight into the nature of this type of consonant variation than resorting to categorical rules of epenthesis and assimilation. Surprisingly, Bradley seems to miss this generalization when he questions the adequacy of the gestural account to explain why assibilation doesn't result from homorganic clusters. Bradley then provides a phonetically-motivated OT analysis of the phenomenon , which makes for an interesting exercise but unfortunately does not add much insight into the nature of the process.
Manuel Díaz-Campos, Acquisition of sociolinguistic variables in Spanish: Do children acquire individual lexical forms or variable rules? Manuel Díaz-Campos contributes to this volume with a study about the acquisition of sociolinguistic variables by Venezuelan children. He investigates whether children of different age groups have acquired the stylistic and sociolinguistic variation that is associated with intervocalic /d/ deletion in Venezuelan Spanish. One of the goals of the study is to test whether acquisition of variation is done on a case- by-case basis or, more generally, by rule. The study includes a series of independent variables or factors such as the dictionary frequency of the lexical items used in the design, the frequency within the corpus and the age group of the participants. The dependent variable is the presence vs. absence of alternation in the pronunciation of intervocalic /d/. Even though the results are presented in a somewhat confusing manner-the interaction between the age variable and the frequency variables is not quite explained—the outcome of the experiments show that the factors having to do with frequency (dictionary or corpus) play a crucial role in predicting the pronunciation of intervocalic /d/, whereas age does not seem to be a significant factor. The weight of the frequency variables is taken by the author to support the hypothesis that variation is not acquired by rule, but on a case-by-case basis.
Marta Ortega-Llebaria, Interplay between phonetic and inventory constraints in the degree of spirantization of voiced stops: Comparing intervocalic /b/ and intervocalic /g/ in Spanish and English. The last paper in this collection presents a study by Marta Ortega- Llebaria in which the author looks for evidence of the theoretical assumption that a particular phonological inventory imposes constraints on the phonetic realization of specific vowels and consonants. She compares the degree of lenition in intervocalic /b/ and /g/ in Spanish and English. Since English distinguishes phonologically between /b/ and /v/ while Spanish has only /b/, it is hypothesized that Spanish will allow for more variability, i.e., more lenition, in the production of /b/ than English. On the other hand, since neither language contrasts /g/ with a voiced velar fricative, both should exhibit similar degrees of lenition in /g/. The results of an acoustic study show that, indeed, Spanish /b/ exhibits more lenition than English /b/, while /g/ shows more similar degrees of lenition in both languages. In spite of the clear results, two considerations can be made regarding this study. First, treating voiced stop lenition as essentially the same phenomenon in Spanish and English, though probably appropriate from an articulatory point of view, ignores the fact that in Spanish this is a generalized phonological process, whereas in English it could be considered a reduction typical of casual speech. Second, the more variable realization of /g/ could also be related, on the one hand, to the specific nature of tongue-body kinematics and, on the other hand, to the difficulties in sustaining voicing in velar stops, which has been proven to cause incomplete stop closures in a variety of languages.
Taken as a whole, this volume is a welcome contribution to the field of laboratory phonology and, in particular, for those of us who are interested in the phonetics/phonology interface of Spanish. I believe the editor needs to be congratulated not only for having put this volume together, but also for kick-starting the series of conferences on Spanish laboratory phonology, now heading towards its third edition.
The selection of papers is representative of a wide variety of interests within the general field of the phonetics/phonology interface, touching upon well-known segmental and suprasegmental issues, but also providing novel data from acquisition and variation studies. One possible criticism as far as the selection of papers is concerned could be the excessive concentration on prosody, with seven out of the ten papers dealing with prosodic issues. This might simply be a fair representation of the papers that were presented at the conference, but perhaps a more balanced distribution between segmental and suprasegmental issues might have been more appropriate for a volume of these characteristics. Also, the three parts in which the volume is divided are not very well balanced, especially part 2, where three of the four papers are about essentially the same thing and draw on data from each other excessively. Perhaps this is due to the editor's personal interests, but it seems a shame that other papers were left out. On the other hand, part 2 presents an in-depth treatment of the relationship between stress and syllable structure, but again this might not be the best philosophy for this type of volume.
From a typographical point of view, the book is reasonably free of typos and other editorial mistakes. There are, however, a few errors that are confusing for the reader, since they detract from a straightforward understanding of the text. For example, in Lleó, Rakow and Kehoe's paper, the explanation of Figure 12 makes reference to the intonational pattern of the word 'Wasser' (in 'Wasserflugzeug') but the figure shows only the second part of the word i.e., 'Flugzeug'. In Prieto's paper, Figure 7 appears as belonging to speaker GH but the text refers to it as MN. Also, in Bradley's paper, the introduction to Section 5 refers to the coming sections as 4.1, 4.2 and 4.3 instead of 5.1, 5.2 and 5.3, respectively.
I would like to make one final observation regarding the use of experimental methods and the presentation of data in some of the papers in this volume. As is unfortunately rather common in some of the work currently being done on laboratory phonology, it seems that some times the experimental data is viewed simply as a complement to the theoretical phonological analysis. Because of this, there can be a certain lack of rigor in the gathering, analysis and presentation of experimental results. Some of these problems have been pointed out above with respect to the specific papers. In my opinion, this lack of experimental rigor can in some cases undermine the theoretical statements that are based on the experimental data. In other instances, as in those papers that present OT analyses of the data, the distance between the experimental results and the assumptions made in the theoretical analysis represents such a leap of faith and such a simplification of the physical reality that one wonders why the authors bothered to gather experimental data at all. I believe two of the pending issues in some laboratory phonology work are the excessive dependence on formalism and the weakness in the treatment of experimental data. The continuation of series like the one initiated with this volume are the best guarantee that these issues will be resolved in future work.
Face, Timothy L. (2000) The role of syllable weight in the perception of Spanish stress. Hispanic Linguistics at the Turn of the Millennium, ed. by Héctor Campos, Elena Herburger, Alfonso Morales-Font and Thomas J. Walsh, 1-13. Somerville, MA: Cascadilla Press.
ABOUT THE REVIEWER:
ABOUT THE REVIEWER
Joaquín Romero is an Associate Professor in the Department of Anglogermanic Philology at the Universitat Rovira i Virgili, in Tarragona, Spain. He received his Ph.D. in Linguistics from the University of Connecticut in 1995. He is currently a research affiliate at Haskins Laboratories, in New Haven, Connecticut. He is interested in speech production and in the phonetics/phonology interface of segmental issues, with special concentration on the relevance of temporal aspects in processes of assimilation and articulatory reduction.