2. Content

In chapter 2, Hendriks provides an overview of examples of asymmetries in L1 acquisition. These examples are not limited to any particular subcomponent of grammar, and are found in morphosyntax, syntax, semantics, and pragmatics. Here she makes the case for the Language-as-Signal approach that employs ranked, violable constraints, just like OT. From the standpoint of production, children prior to age 5 may not have yet acquired the proper ranking of constraints. As far as perception is concerned, it may be the case that children are not yet able to take into account the perspective of the speaker. Asymmetries where production precedes comprehension and vice versa are expected, since prior to age 4 it is generally assumed that children have not yet acquired either the proper ranking of constraints in production nor an ability to take into account the speaker's perspective (comprehension). Agreeing with Hendriks, the pattern of production before comprehension is surprising and difficult to explain within the Language-as-Code approach.

Chapters 3 and 4 expound on anaphor/pronoun resolution asym-metries in 5-year-old Dutch- and English-speaking children. The listener's perspective is the focus of chapter 3, whereas in chapter 4, the emphasis is on the speaker's perspective. Hendriks points out that errors in the comprehension of object pronouns are still common for Dutch- and English-speaking children even after they have started producing these same forms. In chapter 3, Hendriks reviews and compares four purported explanations for these asymmetries: the pragmatic explanation, the grammatical explanation, the task-based explanation, and the interface explanation. Hendriks argues for the grammatical approach, stating that this line of reasoning is consistent with an OT-grammar in explaining Delay of Principle B effects. To adduce further support for the grammatical explanation for these production/comprehension asym-metries, Hendriks models the findings from her experimental data discussed in chapters 3 and 4 in the cognitive architecture ACT-R (Anderson Reference Anderson2007). She argues that computationally modeling the grammatical explanation within ACT-R “serves as a proof of concept that the bidirectional approach is viable” (p. 111).

Turning to the speaker's perspective in chapter 4, Hendricks high-lights the tendency of children at or prior to ages 4–6 to (over)use ambiguous pronouns due to their inability to consider the listener's perspective (due possibly to a lack of a fully-developed Theory of Mind in children at this age).Footnote ¹ Looking at the production of pronouns combined with word order constrained by discourse requirements, Hendriks demonstrates how the same constraints used in chapter 3 to account for the listener's perspective can be used to model these production asym-metries.

The fact that production/comprehension asymmetries are not universal crosslinguistically is the focus of chapter 5. Of particular interest is an observation that Germanic languages such as Dutch and English show Delay of Principle B effects, whereas Romance languages apparently do not. From an OT-perspective, one way of accounting for these crosslinguistic differences is by postulating that this contrastive behavior boils down to different constraint rankings. As noted in section 5.4, another important factor affecting the lack of Delay of Principle B effects in Romance languages concerns the structural properties and lexical inventories of these grammars. For example, Romance languages also have clitic pronouns (whereas most Germanic languages and dialects do not), which could be an additional confounding factor for these differences.Footnote ² In chapter 6, Hendriks again addresses the strengths and weaknesses of competing approaches to modeling production/ comprehension asymmetries in language.

3. Concluding Remarks

Hendriks's analysis of production/comprehension asymmetries is highly laudable, as it provides a gateway where formal theorizing and experi-mental approaches to language can effectively intersect. The book is highly accessible to a wide array of scholars, which is certainly not an easy task. Given the growing interest in biOT as a viable framework to model language development and change (Benz & Mattausch Reference Benz and Jason2011, Legendre et al. Reference Legendre, Michael, Henriëtte and Erin2016), this research contributes to and enriches multiple academic communities. Furthermore, this application of biOT in emerging L1 grammars can also be extended to production/compre-hension asymmetries in elderly populations (Hendriks et al. Reference Hendriks, Charlotte and John2014) and language change (Hendriks & van Rij Reference Hendriks and Jacolien2011).

The only aspect of the design of the book that I found troubling was the fact that the theoretical treatment of Optimality Theory—in particular, the way in which bidirectional optimization was defined and used throughout this book—was separated from the main body of the text and appears in the final chapter (appendix: chapter 7). For those interested in the theoretical component of Hendriks's study, the flipping back and forth can become a bit cumbersome. This minor caveat notwithstanding, Petra Hendriks's contribution to understanding asym-metries between language production and comprehension and providing the sketch of a framework for addressing production/comprehension asymmetries represents a substantial contribution to the fields of psycholinguistics and linguistic theory.