Since Ross's (1967) seminal dissertation on constraints on long-distance syntactic extraction, linguists have attempted to find adequate ways to model constituent displacement in grammatical formalisms. Essential questions concerning mobility, that is, which constituents can be ex-tracted, and permeability, that is, out of which particular syntactic contexts individual lexical items can be extracted, have guided research in long-distance extraction since the inception of generative grammar.

However, as Kiziak points out, many of the supposed “answers” present in contemporary theoretical models—especially those concerning German data—are not uncontroversial. In this monograph, which repre-sents a significantly revised version of her dissertation, Tanja Kiziak takes a closer look at subject/object extraction asymmetries in German. The objective of Kiziak's study is to use her findings on subject/object asymmetries in German to determine whether subject positions in German are internal or external to the VP, and to weigh in on the “paren-thesis versus extraction” debate (that is, do these constructions involve some form of extraction, or do they appear base-generated as an entire phrase?) concerning purported instances of extraction from verb-second clauses. Kiziak's contribution to this theoretical debate is a study with a strong empirical basis, which also allows her to identify some of the underlying causal factors for patterns observed in the judgment studies conducted here.

The judgment tasks were presented as online questionnaires, carried out via the WebExp software package. Both magnitude estimation as well as thermometer judgments were used to statistically determine the elicited judgment values. The experiments presented in this monograph tested the contribution of the following four factors determining the permissibility of long-distance extraction in German: 1) mobility, 2) permeability, 3) bridge quality (certain verbs license long-distance extraction better than others), and 4) movement type (wh-movement versus topicalization).

Chapters 3 through 6 present a series of judgment studies. In chapter 3, Kiziak focuses on previous arguments in the literature and introduces new experimental data related to the subject/object extraction asym-metries in German. She discusses three judgment studies that involve long-distance wh-movement and long-distance topicalization from dass-clauses, ob-clauses, and embedded wh-questions. What emerges from these experiments is that subject/object extraction asymmetries do exist in German, although they are obscured at times by other factors (for example, nominative case-marking on the extracted subject and possibly the generative constraint known as the Empty Category Principle (ECP); Chomsky 1981).

The experiments introduced and discussed in chapter 4 tackle issues pertaining to long extraction from embedded verb-second clauses versus parenthetic construction debate. These experimental results support the perhaps controversial conclusion that these structures are best classified as parenthetic constructions (although Kiziak is hesitant to claim that the extraction analysis is “falsified” by these findings, p. 121).

In chapters 5 and 6, Kiziak discusses previous contrasting research, which locates the subject/object extraction asymmetry in the matrix clause and those studies, in contrast, which locate the explanation for these asymmetries in the embedded clause. As discussed in chapter 5, one of the key factors that influences long-distance extraction in German is the case marking on the extracted DP (although Kiziak is quick to point out that this factor is not a “fix-all” in that case marking cannot by itself explain the data gathered from the experiments discussed in chapter 3). Due to the established differences between the case systems of English and German, the difference observed with respect to traditional that-trace effects are not as prevalent in German as they are in (standard) English.

Kiziak investigates surface conditions of the embedded clause and the properties of the subject extraction site in chapter 6. One key finding from these experiments reveals that subject extractions are degraded if they violate the that-trace filter. In addition to this finding, Kiziak argues that “if the violation was disguised by means of a case-ambiguous DP in the embedded clause, ratings for subject extractions were enhanced” (p. 227). Atypical, object-like subjects also proved to be more mobile than prototypical subjects. As a result, Kiziak's findings support extraction theories that refer to properties of the trace/slash-termination. A sum-mary of findings and directions of future research are presented in chapter 7.

In many respects, Kiziak's study is indeed a breath of fresh air in an era where models of theoretical syntax often lose sight of the importance of natural data in pursuit of (often esoteric) theory-internal goals. Since Featherson's (2001) treatment of empty categories from a hybrid theoretical-experimental approach, research following in his footsteps has advanced our understanding of natural language syntax rather than hindered it. The findings presented here alongside a detailed, well-conceived empirical research program contribute a wealth of under-standing concerning the subject/object extraction asymmetry debate in German and beyond.

The only critical remark on an otherwise brilliant piece of research is the absence of any detailed treatment of how Minimalism (the theoretical model employed by Kiziak) could and should be adjusted to capture these empirical findings. For example, if, as the results indicate, theories of extraction should account for properties of the trace position, what implications would this have for phase-theory within modern Mini-malism? Although there are occasional allusions to necessary theoretical adjustments throughout the manuscript, they appear infrequently and do not constitute a coherent portion of the text. Regardless of this minor shortcoming, which I hope the author will take up in related work in the near future, Kiziak's study should serve as a model for a promising research program that successfully merges empirical and theoretical studies of language.