Symbolic and automatic differentiation of languages

Formal languages are usually defined in terms of set theory. Choosing type theory instead gives us languages as type-level predicates over strings. Applying a language to a string yields a type whose elements are language membership proofs describing how a string parses in the language. The usual building blocks of languages (including union, concatenation, and Kleene closure) have precise and compelling specifications uncomplicated by operational strategies and are easily generalized to a few general domain-transforming and codomain-transforming operations on predicates. A simple characterization of languages (and indeed functions from lists to any type) captures the essential idea behind language “differentiation” as used for recognizing languages, leading to a collection of lemmas about type-level predicates. These lemmas are the heart of two dual parsing implementations—using (inductive) regular expressions and (coinductive) tries—each containing the same code but in dual arrangements (with representation and primitive operations trading places). The regular expression version corresponds to symbolic differentiation, while the trie version corresponds to automatic differentiation. The relatively easy-to-prove properties of type-level languages transfer almost effortlessly to the decidable implementations. In particular, despite the inductive and coinductive nature of regular expressions and tries respectively, we need neither inductive nor coinductive/bisimulation arguments to prove algebraic properties.

Neural Evidence of Language Membership Control in Bilingual Word Recognition: An fMRI Study of Cognate Processing in Chinese–Japanese Bilinguals

This study aims to examine the neural mechanisms of resolving response competition during bilingual word recognition in the context of language intermixing. During fMRI scanning, Chinese–Japanese unbalanced bilinguals were required to perform a second-language (L2) lexical decision task composed of cognates, interlingual homographs, matched control words from both Chinese (first language) and Japanese (L2), and pseudowords. Cognate word processing showed longer reaction times and greater activation in the supplementary motor area (SMA) than L2 control word processing. In light of the orthographic and semantic overlap of cognates, these results reflect the cognitive processing involved in resolving response conflicts enhanced by the language membership of non-target language during bilingual word recognition. A significant effect of L2 proficiency was also observed only in the SMA, which is associated with the task decision system. This finding supports the bottom-up process in the BIA+ model and the Multilink model. The task/decision system receives the information from the word identification system, making appropriate responses during bilingual word recognition.

EAL Writers and Peer Tutors: Pedagogies that Resist the “Broken Writer” Myth

Writing centres offer a safe space for writers, including English-as-additional-language (EAL) students, to negotiate meaning and become more <luent with academic writing genres. However, a disconnect still exists between the writer-centred principles that inform WC tutoring practice and the pervasive myth that writing centres repair “broken” writing. An analysis of data from a writing centre’s client reports, as well as peer tutors’ comments and student writing samples, indicates that a student’s language membership does not predict types of writing challenges or errors. This <inding inspired a roundtable discussion about pedagogical approaches that not only empower EAL students but help writing centres resist the “broken writer” myth.

Language membership as a gradient emergent feature

In their keynote article, Dijkstra, Wahl, Buytenhuijs, van Halem, Al-jibouri, de Korte, and Rekké (2018) introduce the Multilink model representing an integrated bi/multilingual lexicon. This proposal builds upon both previous and recent research on an integrated cognitive architecture underlying the language faculty (for a summary, see e.g., Putnam, Carlson &amp; Reitter, 2018). In our view, the adjustments proposed by the authors are an improvement on previous instantiations of similar models such as those discussed in the present article. In our remarks we explicate how the Multilink model may be further enhanced, by making any appeal to language-specific nodes or representations epiphenomenal. To achieve this, we propose a novel approach to representing language membership as the result of gradient emergent principles that builds upon the integrated lexicon underlying the Multilink model.

On keeping cool: The role of inhibition in bilingual word processing

One of the cool aspects of the original implementation of the BIA model (van Heuven, Dijkstra &amp; Grainger, 1998) was the discovery that inhibitory connections between language nodes and lexical representations was a necessary feature for the model to be able to simulate the target data set at that time. This demonstrates the importance of computational modeling, a key point of the present target article, since inhibitory connections were postulated to occur only between representations at the same level in the conceptual model (Grainger &amp; Dijkstra, 1992). Top-down inhibition was subsequently dropped in the BIA+ model (Dijkstra &amp; van Heuven, 2002), and the Multilink model of the present target article (Dijkstra, Wahl, Buytenhuijs, van Halem, Al-jibouri, de Korte &amp; Rekké, 2018) goes one step further by removing all kinds of inhibitory connections, both between and within levels. Instead, the authors of the model propose that bilingual language processing relies on bidirectional excitatory connections between representations at different levels. This is curious given that even more evidence has accumulated in favor of inhibition since the original implementation of the BIA model, both between neighboring lexical representations (i.e., lateral inhibition) and from language membership representations (e.g., language nodes and tags) down to lexical representations. In this commentary, we focus on whether the exclusion of these two inhibitory processes is warranted, and how the inclusion of these processes might benefit future developments of the model.

When a seven is not a seven: Self-ratings of bilingual language proficiency differ between and within language populations

Self-ratings of language proficiency are ubiquitous in research on bilingualism, but little is known about their validity, especially when the same scale is used across different types of bilinguals. Self-ratings and picture naming data from 1044 Spanish–English and 519 Chinese–English bilinguals were analyzed in five between- and within-population comparisons. Chinese–English bilinguals scored more extremely than Spanish–English bilinguals, and in opposite directions at different endpoints of the self-ratings scale. Regrouping bilinguals by dominant language, instead of language membership, reduced discrepancies but significant group differences remained. Population differences appeared even in English, though this language is shared between populations. These results demonstrate significant problems with self-ratings, especially when comparing bilinguals of different language combinations; and subgroups of bilinguals who speak the same languages but vary in acquisition history and/or dominance. Objective proficiency measures (e.g., picture naming or proficiency interviews) are superior to self-ratings, to maximize classification accuracy and consistency across studies.

Cross-language semantic influences in different script bilinguals

The current study examined automatic activation and semantic influences from the non-target language of different-script bilinguals during visual word processing. Thirty-four Arabic–Hebrew bilinguals and 34 native Hebrew controls performed a semantic relatedness task on visually presented Hebrew word pairs. In one type of critical trials, cognate primes between Arabic and Hebrew preceded related Hebrew target words. In a second type, false-cognate primes preceded Hebrew targets related to the Arabic meaning (but not the Hebrew meaning) of the false-cognate. Although Hebrew orthography is a fully reliable cue of language membership, facilitation on cognate trials and interference on false-cognate trials were observed for Arabic–Hebrew bilinguals. The activation of the non-target language was sufficient to influence participants’ semantic decisions in the target language, demonstrating simultaneous activation of both languages even for different-script bilinguals in a single language context. To discuss the findings we refine existing models of bilingual processing to accommodate different-script bilinguals.

Evaluating the predictions of three syntactic frameworks for mixed determiner–noun constructions

This paper presents a comparative evaluation of three linguistic frameworks, the Minimalist Programme (MP), Word Grammar (WG) and the Matrix Language Frame Model (MLF), regarding their predictions of possible combinations in a corpus of 187 German–English code-switched (CS) determiner–noun constructions. The comparison revealed a significant difference in the accuracy of the predictions between the MP and WG, but not between the other frameworks. We draw attention to the fact that while WG and MP deal with the processes of feature agreement between determiner and noun, the MLF is concerned with a broader notion of agreement in language membership. We suggest that advances in our understanding of grammaticality in code-switching will be achieved by combining the insights of all three frameworks instead of considering them in isolation.