Book review

Split Auxiliary Systems

Aranovich, Raúl (ed.) 2006. Split Auxiliary Systems. 2006. Amsterdam and Philadelphia: John Benjamins.

Reviewed by Wolfgang Schulze, Institute for General Linguistics and Language Typology, Ludwigs-Maximilians-Universität München

Split Auxiliary Systems (SAS) is a collection of articles resulting from a workshop on "Cross-linguistic Variation in Auxiliary Selection" at the University of California, Davis (May 31-June 1, 2003). Basically, Split Auxiliary Systems concern the alternation between auxiliaries with the meanings of BE and HAVE. Depending on the individual language, there may be two or more construction types differing in the use of the corresponding auxiliary. The alternation can be related to a number of functional and semantic properties of the lexical verb, for instance to transitivity or variation with respect to the control scale.

The Language Index of the SAS volume illustrates that Split Auxiliary phenomena are frequently described synchronically or diachronically for Romance and Germanic languages. The index lists Italian, Italian dialects, French including QuC)bec French, Spanish, Romanian, Catalan and Latin itself for the Romance languages, and Danish, Dutch, English, German, Icelandic, and Old Saxon for Germanic. Beyond that, Chinese and Korean data are taken into consideration. The fact that the categorical domains effected by an SAS mainly concern the formation of Past tense and Perfect forms in both Germanic and Romance (plus aspects of reflexivity in Romance) suggests that the most of the types of SAS discussed in the volume reflect some kind of sprachbund phenomenon that is based, however, on more general principles of subcategorizing mental event images.

The SAS volume (277 pages in sum) comprises eleven articles followed by a language index, a name index, and a subject index. The individual articles are arranged in alphabetic order of the names of their authors. Hence, the volume does not aim to be a reference book on SAS. Rather, the reader finds a number of case studies that illustrate the formal and functional range of SAS phenomena, especially in Western European languages (plus Chinese and Korean). The articles meet the high standard of a handbook; the volume incorporates some variety as the individual authors subscribe to different theoretical frameworks. The reader thus has to accommodate to theoretical positions that are sometimes in conflict, ranging from formal approaches via Optimality Theory to Functionalism.

The volume starts with an account of "Split Auxiliary Selection from a Cross-Linguistic Perspective" by the editor Raúl Aranovich. He gives a brief overview of the different approaches proposed to SAS so far, paying special attention the significance of SAS for the Unaccusative Hypothesis. The author also surveys several options to pinpoint the motivation of SAS phenomena arguing that they should be arranged according to an Auxiliary Section Hierarchy that runs from "directed change of location" verbs (BE) to "controlled volitional activities (HAVE). In addition, Aranovich opts not to interpret SAS as just a diagnostic tool to test Unaccusativity, but to analyze the phenomenon as a syntactic and semantic constructional strategy for its own that is also governed by hierarchic features of referential items and subjected to diachronic change.

Varlee Arnett's article on "Transitivity and Auxiliary Selection by L2 Students of German" addresses the question of whether the way English speakers acquire the German SAS paradigm can help us to better understand how the selection of auxiliaries is determined by the "gradient model", related to the Auxiliary Selection Hierarchy. Arnett carefully parameterizes the learning behavior of a number of test persons and concludes that the feature "change of location/state" (mutative in terms of the author) represents the prototypical feature to be associated with a BE-patterning. However, this feature is strongly correlated with aspects of overt transitivity, that is with the question how many participants become relevant in the construction of (in my terms) an event image.

The next article turns to "An Irrealis BE Auxiliary in Romanian" (Larisa Avram and Virginia Hill). Romanian is a strong HAVE-language with respect to unaccusativity, but occasionally uses the auxiliary fi (BE) in this respect. In addition, uninflected fi is used in the irrealis mood, where the HAVE-paradigm (am, ai, a etc.) is not possible at all. The authors thus show that SAS is not necessarily confined to the typical domains of the Unaccusative Hypothesis. The analysis is carried out using the Minimalist Program and also considers differences in word order and agreement.

Paduan, a North-Eastern Italian dialect is addressed in Michaela Cennamo's and Antonella Sorace's paper on "Auxiliary Selection and Split Intransitivity in Paduan: Variation and Lexical-Aspectual Constraints" In Paduan, the cut-off point in the Auxiliary Selection Hierarchy is shifting towards the left, that is, more verbs tend to select HAVE than in Standard Italian. Even some verbs quite at the endpoint of the scale (change-of-state-verbs) take the HAVE option. The authors test several options to explain this phenomenon from a syntactic point of view, but arrive at the conclusion that the internal semantics of the verb as well as aspectual features mainly govern the selection of auxiliaries in Paduan.

A challenging view is presented in Bridget Dinka's contribution "The development of the HAVE Perfect: Mutual Influence of Greek and Latin". In fact, the paper should have been placed at the very beginning of the volume, because it describes the emergence of the Latin HAVE construction that again lays the ground for the multiple types of SAS phenomena in Romance languages. The author argues that the Latin HAVE construction has resulted from loan translations of Greek texts in times of Plautus, Cato the Elder and others. In a second step, the Latin construction was projected back onto Koine Greek marked for the merger of Aorist and Perfect tense forms. As Koine was strongly identified with the language of the New Testament and thus with Christianity, the HAVE construction is said to have become a stylistic indicator of Christian texts both in Greek and Latin, which would explain its 'propagation' in the later varieties of Latin and in Romance language.

In her article "Agentivity versus auxiliary choice", Vera Lee-Schoenfeld relates SAS phenomena to the accusativum cum infinitivo-construction of German. The phrase-structural account concentrates on the question of whether auxiliary selection can be used as a diagnostic tool for unaccusativity/unergativity in German. Instead of explaining AcI constructions in terms of the SAS pattern of German, the author argues for an account based on pronominal binding.

The Unaccusative Hypothesis is also considered in Géraldine Legendre's contribution on "Optimizing Auxiliary Selection in Romance". In addition, the author interprets the Auxiliary Selection Hierarchy in terms of Optimality Theory introducing a set of violable constraints on the use of BE and HAVE patterns. She argues that OT not only accounts for the differences between Italian and French concerning the cut-off point on the BE/HAVE-scale, but also for the diachronic processes that underlie the current distributional patterns.

Feng-hsi Liu's article on "Auxiliary Selection in Chinese" is one of the two articles in the volume that explicitly address non-European languages. The author analyzes the two Chinese aspectual auxiliaries -le and -zhe in connection with the Auxiliary Selection Hierarchy, positing that the features telic/atelic and dynamic/stative mark the endpoints of this scale. The selection of -le and -zhe is in addition determined by the semantics of the event image, more concretely by features of agentivity.

In her contribution "Parameterized Auxiliary Selection", Janet H. Randall uses mainly German and Dutch data to propose an alternative to the Auxiliary Selection Hierarchy as a solution to explain SAS phenomena within and across languages. This alternative is termed "Parameterized Linking" and relates an Argument Structure to the level of Conceptual Structure. Parameterized Linking aims at lexicon-syntactic mapping and considers more broadly sentence-level properties of verbs parameterized according to a universal ensemble of linking procedures.

Seongha Rhee's article on "Particle Selection in Korean Auxiliary Formation" strongly opts for a more pronounced consideration of particle-like forms as the source for grammaticalized elements, here for the varying patterns of the issta auxiliary. The author carefully describes the role of basically three particles (-e (sequential), -ko (focusing), and -key (manner/purpose) in the development of the paradigm of the issta auxiliary (e-iss-, -ko-iss-, -key-iss- etc.).

Finally, K. Aaron Smith discusses "Language Use and Auxiliary Selection in the Perfect". Contrary to the other articles in the volume, Smith starts from a usage-based model of language storing and processing to explain the differences of SAS phenomena in English and German. In addition, the author convincingly argues for the integration of corpus-based data from diachronically different stages of a language to illustrate the way SAS patterns become entrenched. The author reveals a set of individual conditions (textual types, grammatical features, frequency etc.) that governed the way how the Germanic BE/HAVE-pattern has singled out in the two languages.

Evaluation

Split Auxiliary Systems is an interesting collection of articles that undoubtedly helps to better understand phenomena related to auxiliary selection. The book is well-edited and nearly free from typographical errors (one exception is p.13, ex. 19: habben should read haben).

The individual contributions offer both a wealth of data and thoughtful analyses, which however are not always comparable to due to the fact that the authors employ different theoretical frameworks. Nevertheless, assumptions related to the Auxiliary Selection Hierarchy as pronounced by Sorace (2000) serve as a leitmotif for most of the contributions, just as frequent references towards "unaccusativity" and "unergativity" illustrate a strong commitment towards formal syntactic approaches. Sometimes, the reader may even be left with the impression that the discussion of SAS phenomena mainly aims at the testing or elaboration of the Unaccusative Hypothesis (interestingly enough, none of the papers refers to Kuno and Takami (2004) as a major contribution for testing this hypothesis). The fact that many of the articles refer to formal syntax (or its derivatives) as an explanatory device results in a somewhat unbalanced presentation of options to explain SAS phenomena. Three major exceptions are Dinka's, Rhee's, and Smith's contributions that show alternative ways to approach the explanation of SAS patterns. The strong formal syntactic orientation of some of the papers results in sometimes problematic claims. For instance, Lee-Schoenfeld makes extensive use of German analytic causative constructions to illustrate the AcI. However, it may well be questioned whether such constructions are AcI-constructions at all, as they can be interpreted instead as a constructional type that also encompasses modal constructions.

One might perhaps make the criticism that most of the articles dwell upon Germanic and Romance data. In fact, SAS phenomena are reported in numerous languages around the world, and it would be a challenging task to see whether the motivation of the Auxiliary Selection Hierarchy often alluded to in the individual papers accounts for related split phenomena in other languages. On the other hand, the heavy orientation towards Germanic and Romance data has two advantages: First, the data of the individual articles become comparable and thus nicely illustrate the differences in explaining the findings. Second, the extremely difficult task of pinpointing the semantic and syntactic motivations of SAS phenomena calls for a good command of the languages at issue and for a careful presentation of the data. The fact that both Germanic and Romance are well-documented guarantee that the subtle "tonality" of split phenomena especially in the 'middle' of the Auxiliary Selection Hierarchy can be adequately described. Nevertheless, it should be added that I am not always sure about the overall quality of the data. For instance, Lee-Schoenfeld gives the German example (p. 123) Hans lässt den Stein sich auf dem Kopf fallen 'Hans lets the rock fall on his head'. However, in Standard German, this phrase is odd as to the position of the reflexive pronoun: German, being a language with a heavy pronominal left shift, would prefer Hans lässt sich den Stein auf den Kopf fallen.

The preference for discussing the SAS typology with the help of data from Romance and German finally evokes another set of questions rarely addressed except in Dinka's paper: Why do German and Romance share, at least historically, a typology that is alien to Slavic, but then again persistent in Slavic-influenced Yiddish? What has been the role of interlinearizing translation from Latin? Why does the SAS typology concentrate on Italian, Continental German, and French, but is weaker for instance in Spanish and Present-Day English? Why did e.g. Terracinese Italian level the SAS pattern to BE (resulting in pattern quite similar to Slavic, Armenian etc.)? What is the role of language contact? We cannot expect the volume to give comprehensive answers to all these questions (as well as to the question how Cognitive Linguistics / Cognitive Grammar(s) would interpret the phenomenon). However, the SAS volume is a first, indispensable step to tackle the problem from a broader perspective.

References

Kuno, Susumo and Ken-ichi Takami. 2004. Functional Constraints in Grammar. Amsterdam/Philadelphia: John Benjamins.

Sorace, Antonella. 2000. Gradients in Auxiliary Selection with Intransitive Verbs. Language 76:859-890.

Links

© 2002-present ICLA; all rights reserved.
Site admin information and contact