EDITORS: Wratil, Melani; Gallman, Peter TITLE: Null Pronouns SERIES TITLE: Studies in Generative Grammar [SGG] 106 PUBLISHER: De Gruyter Mouton YEAR: 2011
Egor Tsedryk, Department of Modern Languages and Classics, Saint Mary’s University, Halifax, Canada
The concept of null subject pronouns has been a hot topic in generative linguistics since the emergence of the parametric approach to the study of grammar started taking its first steps in the late seventies. Whether or not a language allows dropping a referential pronominal subject in finite clauses (i.e. pro-drop) is, in fact, a classical example of a parameter; it has either a positive or a negative value, which is fixed during language acquisition. This volume is one of the recent attempts to revisit the pro-drop parameter through both synchronic and diachronic perspectives, with fairly wide empirical coverage, including Germanic, Celtic, Romance, Finno-Ugric and Creole languages.
Though not explicitly stated by the editors, I infer from the footnotes that this book represents a collection of five papers that appear to be selected from the workshop “Nullpronomen” (University of Jena, September 2006). The introduction provides a brief historical background with regard to two major issues. The first concerns the status of null pronouns in general and null subjects of finite clauses in particular. The basic question is whether they all belong to the same empty category known as ‘pro’ in the framework of Principles and Parameters. The second issue concerns the nature of the pro-drop parameter. Under standard assumptions, its binary setting should be inferred from an independent linguistic property (e.g. richness of verbal inflection). However, according to the editors, the volume’s contributors seem to converge on the idea that there is no general syntactic or morphological property directly responsible for pro-drop phenomena cross-linguistically.
The volume is organized as follows: The first three papers deal with null subject phenomena related to historical, as well as synchronic changes occurring with verb placement, clitics and verbal agreement. Moreover, agreement-related null subjects are opposed to those that are discourse-related. The last two papers shift the focus from null subjects of finite clauses to null possessors of possessive constructions. I will first summarize each paper individually and then evaluate the book as a whole.
In “Pro-drop in the history of German -- From Old High German to the modern dialects,” Katrin Axel and Helmut Weiß focus on the difference observed between Old High German and modern Continental West-Germanic dialects with regard to pro-drop properties. In the former, pro is licensed predominantly in root sentences with the verb in second position (V2), whereas, in the latter, pro-drop is limited to subordinate clauses featuring complementizer agreement. The authors argue that this difference is only superficial, and that the underlying configuration for pro-drop sentences remained the same; ‘pro’ must be c-commanded by Agr (i.e. an agreement morpheme). The only restriction that applies in modern dialects is that Agr must be pronominal in order to license ‘pro’. According to the authors, pronominal Agr originates from clitic constructions; a subject clitic that immediately follows the inverted verb (e.g. in V2 contexts) gradually becomes an inflectional suffix. Then, complementizer-clitic constructions are reanalyzed in the same way by analogical extension. This is reflected in double agreement systems, in which the choice of verbal inflection depends on the syntactic position of the verb (i.e. first, second or clause-final). For example, in central Bavarian, the verbal inflection in V2 sentences is not the same as those in verb-final and, most importantly, is identical to complementizer agreement.
In “Historical pathways to null subjects: Implications for the theory of pro-drop,” Eric Fuß describes two different pathways of pro-drop development: reanalysis of subject clitics as agreement markers, on one hand, and emergence of discourse-oriented pro-drop in Creole languages, on the other. With regard to these clitics, Fuß shows that, while enriching the verbal inflectional paradigm, their reanalysis creates gaps in their own paradigm. Such gaps trigger deblocking of a universal zero spell-out rule (i.e. null realization of a clitic pronoun) at the point of post-syntactic vocabulary insertion (in terms of Distributed Morphology). More precisely, if no overt clitic competes for morphosyntactic features, a null weak pronoun is inserted into a determiner head at the syntactic level (clitics are assumed to be neither minimal nor maximal projections).
Regarding the example of Bavarian and Non-Standard French, Fuß shows that only less distinctive inflectional markers are replaced by more distinctive pronominal forms (if available), which usually gives rise to partial pro-drop. As an alternative pathway, languages can develop a full pro-drop system based on the discursive identification of null subjects, as in Mauritian Creole and Chabacano (i.e. Spanish-based Creoles spoken in the Philippines). According to the author, these Creoles have been influenced by neighboring Austronesian languages (e.g. Malagasy, Tagalog and Cebuano), whose elaborate voice systems allow dropping topic arguments of different kinds (including objects). As Fuß points out, his analysis does not argue against agreement-related theories of pro-drop, but suggests a more restrictive role of rich agreement in pro-drop phenomena.
In “Uncovered ‘pro’ -- On the development and identification of null subjects,” Melani Wratil proposes that, instead of a hard-wired parameter provided by the universal grammar, pro-drop properties should be considered as a competition between optimization of perception and phonological minimization of pronouns and agreement markers. These two forces produce a grammaticalization cycle, going back and forth from one antipode (i.e. full pro-drop) to another (i.e. full non-pro-drop), with an intermediate state of partial pro-drop. The author illustrates this cycle, showing how erosion of agreement in Old French (full pro-drop) produced a non-null subject language (modern Standard French), while passing through the stage of partial pro-drop in Middle French. At the same time, subject clitics in present-day colloquial varieties of French are gradually reanalyzed as agreement markers becoming an integral part of verbal inflection. In other words, modern Non-Standard French started showing symptoms of a null subject language. In addition, the author points out that two types of ‘pro’ can emerge during the partial pro-drop stage: anaphoric (i.e. locally bound by a verbal agreement marker) and non-anaphoric. Only the latter corresponds with the classical definition of ‘pro’ as a definite pronoun, subject to Principle B of Binding Theory. Wratil shows how these types of ‘pro’ are distributed in Irish, Welsh and Básuse Saramaccan (an Atlantic Creole language). She also emphasizes that the relation between agreement and pro-drop is not always unidirectional. Thus, on the basis of modern Finnish, she demonstrates that the loss of null subjects can precede and consequently induce impoverishment of verbal agreement.
In the last section of her paper, Wratil uses examples from Nez Perce, Mauritian Creole and Old Icelandic to argue that genuine (i.e. non-anaphoric) null pronouns are recovered from discourse rather than from the immediate grammatical environment, including agreement. For this purpose, she makes a clear distinction between discourse-oriented pro-drop, on one hand, and topic deletion, on the other. The latter is an instance of a variable bound by a left-peripheral null operator, and therefore, has a more restrictive syntactic distribution.
In “Silent resumptives in Zurich German possessor relativization,” Martin Salzmann examines relativization, resumption, null pronouns and the structure of possessive determiner phrases (DPs) in Zurich German. Salzmann argues that resumptive relative clauses are derived through base-generation, while gap-relative clauses are derived by movement. Island insensitivity is one of the main arguments in favor of a base-generation analysis. In fact, Salzmann considers resumption as a last resort strategy, when needed, either to circumvent locality violations or to make an oblique case visible. Interestingly, possessor relativization is also not sensitive to islands, even though it does not seem to use any overt resumptive strategy:
Das deet isch de Schüeler, won i geschter [sin Vatter] käne gleert han. that there is the student, COMP I yesterday [his father] got.to.know have ‘Over there is the student whose father I met yesterday.’ (cf. (19b), p. 158)
Note that Salzmann does not take the possessive pronoun (‘sin’) inside the bracketed DP to be a resumptive pronoun. In fact, there are so-called ‘possessor doubling constructions,’ where an overt dative (DAT) possessor can be inserted in front of the possessive pronoun:
dem Schüeler sin Vatter the.DAT student his father ‘the father of this student / this student’s father’ (cf. (20b), p. 159)
Such constructions are found in non-standard varieties of German, but not in Standard German, and are restricted to 3rd person possessors. Salzmann assumes that every time the possessor is not overtly expressed in German, it is realized as a dative ‘pro’ (the only option available in Standard German). This ‘pro’ is, in fact, the silent resumptive pronoun, which allows possessor relative clauses to obviate locality constraints. It is bound by a null operator base-generated as a specifier of the invariant complementizer (‘won’) introducing relative clauses in Zurich German. The residual question, extensively discussed at the end of Salzmann’s paper, concerns the economy of derivation and competition between base-generation and movement.
In “Anti-agreement with subjects and possessors from a typological perspective: A case for null pronouns or for economy?,” Albert Ortmann refutes the necessity of postulating null pronouns to explain so-called ‘anti-agreement’ or ‘plurality splits’ in Hungarian possessive constructions. The pattern can be summarized as follows: the possessed noun does not agree with the plural possessor, unless the latter is pronominal or null; in other words, lexical possessors do not trigger plural agreement on the noun head. This is shown below (3PL = 3rd person plural, 3SG = 3rd person singular).
az ő ház-uk the they house-3PL ‘their house’ (cf. (2b), p. 226)
a ház-uk the house-3PL ‘their house’ (cf. (5c), p. 227)
a nagynéni-k ház-a the aunt-PL house-3SG ‘the aunts’ house’ (cf. (2d), p. 226)
As Ortmann argues, through reference to the examples above, the fact that the first two pattern alike with regard to agreement does not necessarily mean that a null pronoun is present without an overt possessor. He proposes deriving anti-agreement from faithfulness and markedness constraints in the framework of Optimality Theory. The ranking of relevant constraints is based on two harmonically aligned scales: the Plurality Scale and the Definiteness Scale. The former simply gives higher ranking to morphologically more specific forms (e.g. plural marking), while the latter is motivated by different degrees of discursive salience and is rooted in human cognition. Different rankings produce different results cross-linguistically. The author shows how his proposal could be applied to plurality splits observed in subject agreement in Welsh and possessor agreement in Turkish, thus affirming his analysis as descriptively and explanatorily more adequate than a purely syntactic approach postulating ‘pro.’
These readings are very enriching on both empirical and conceptual grounds. Each paper reports high-quality, thorough research that will interest specialists working on null subject properties through diachronic, synchronic and typological perspectives. Collectively, they are not confined to a specific framework and present analyses from different theoretical and methodological perspectives, covering both synchronic and diachronic dimensions of the data presented. This necessarily leads to opposing points-of-view creating the dynamics of a scientific debate. For example, Axel and Weiß emphasize the importance of the complementizer agreement to license null subjects in Germanic dialects, whereas Fuß objects, stating that this is not a sufficient requirement alone, mentioning West Flemish as an example where an elaborate complementizer agreement exists but, nevertheless, does not exhibit any pro-drop properties (fn. 13, p. 64). At the same time, Fuß proposes that null subject properties in Mauritian Creole have been influenced by Malagasy (pp. 82-84), which is subsequently contested by Wratil (p. 131). Finally, Salzmann assumes the existence of ‘pro’ in German possessive DPs as explaining intricate properties of possessor relativization in Zurich German, whereas Ortmann argues against a null possessor in Hungarian, even though nominal Hungarian anti-agreement seems to indicate the opposite. In fact, such points of disagreement make this volume interesting to read and represent one of its main strengths, along with empirical breadth and depth.
In terms of weaknesses, this book could have been better framed in more recent contributions focusing on null subjects of finite clauses (e.g. Fascarelli 2007, Cole 2009, Biberauer et al. 2010). Even though the introduction provides sufficient background for each paper, it falls short in addressing the latest issues surrounding ‘pro’ and the pro-drop parameter. For example, the editors raise the following question: “[D]o all phonologically empty subject pronouns that occur in the finite clauses of full and partial languages belong to a special group of empty categories called ‘pro’ that combine the features [+pronominal], [--anaphoric] and [+definite]?” (p. 14). While they admit that this question is relevant in the context of “the early Principles and Parameters Theory” (p. 15), it does not seem to follow the current debate in light of more recent attempts to redefine ‘pro.’ For Holmberg (2005), ‘pro’ is not a uniform category with a [+definite] feature. In fact, its apparent definiteness (in consistent null subject languages) is a derived property, while its default interpretation (surfacing in partial null-subject languages) is equivalent to that of a generic, 3rd person indefinite pronoun (see Holmberg 2005: 555).
Moreover, ‘pro’ in partial null-subject languages, such as Hebrew, Finnish and Brazilian Portuguese, can be bound (or controlled) by a matrix clause argument, which undermines its classical definition in terms of being [--anaphoric]. As suggested by Shlonsky (2009), this referential dependency could be attributed to the lack of a person feature in ‘pro.’ Null subjects of embedded finite clauses are very briefly discussed by Fuß (pp. 82, 86); they are also mentioned in Wratil’s paper (p. 130), but her discussion does not extend beyond a brief footnote. Would ‘pro’ be [+anaphoric] in embedded finite clauses?
In the same spirit, the editors address the issue of the pro-drop parameter when proposing the question, “Can we be sure that the innate mental abilities of language learners provide a specific parameter, which is binary in its nature, and whose individual fixation during language acquisition determines whether subject null arguments are allowed or forbidden in the target language?” (pp. 14-15). Once again, there is an oversimplification here. As Holmberg (2010) notes, instead of a single parameter, there is more than one parameter involved, which he emphasizes by stating “The Null Subject Parameter (NSP) is often talked about in the singular, even though it is widely recognised that null subjects can be derived in more than one way, and that, therefore, more than one parameter is involved in determining whether subject pronouns can be null or not in a given language” (Holmberg 2010: 88). If more than one binary parameter is present, it is then not surprising that there is more than one morphological and syntactic property that would be involved in licensing null subjects. This is what is corroborated by the papers collected in this volume anyway. On a side note, it is quite surprising that the volume “Parametric Variation: Null Subjects in Minimalist Theory” (Biberauer et al. 2010), listed in Eric Fuß’s references (under Roberts 2010), is not mentioned in the introduction.
Finally, from a purely editorial point-of-view, I have a comment about Martin Salzmann’s paper, which stands out by its length (almost one third of the whole volume). For some reason, it is the only paper that is preceded by an abstract (written in italics), and has as much information in footnotes as in the main text. At times, a short statement in the text is accompanied by three consecutive footnotes presenting several alternative analyses and secondary arguments, which, in most cases, could be discarded for the sake of readability.
These general criticisms do not, by any means, undermine the quality of individual papers. This volume is a valuable contribution to current research on null subjects and null pronouns in general.
Biberauer, Theresa, Anders Holmberg, Ian Roberts, & Michelle Sheehan (eds.). 2010. Parametric variation: Null Subjects in Minimalist Theory. Cambridge: Cambridge University Press.
Cole, Melvyn. 2009. “Null subjects: A reanalysis of data.” Linguistics 47:559-587.
Frascarelli, Mara. 2007. “Subjects, topics and the interpretation of referential ‘pro’: an interface approach to the linking of (null) pronouns.” Natural Language and Linguistic and Linguistic Theory 25: 691-734.
Holmberg, Anders. 2005. “Is there a little ‘pro’? Evidence from Finnish.” Linguistic Inquiry 2005:533-564.
Holmberg, Anders. 2010. “Null subject parameters.” In Biberauer et al. 2010.
Shlonsky, Ur. 2009. “Hebrew as partial-null subject language.” Studia Linguistica 63:133-157.
Roberts, Ian. 2010. “A deletion analysis of null subjects.” In Biberauer et al. 2010.
ABOUT THE REVIEWER
ABOUT THE REVIEWER:
Egor Tsedryk is an Associate Professor of French and Linguistics at Saint
Mary’s University (Halifax, Canada). His research interests include syntax
and theoretical linguistics, and one of his current projects is focused on
null subjects in embedded finite clauses.