In this paper, I propose a formal model of the representation and processing of an integrated multilingual phonological grammar. Two options are considered: (a) a merged union of features from all languages, and (b) an integrated repository which includes representational language tags on hierarchical phonological structures. I argue that the tagged model is preferred given that the merged-union proposal runs into problems with (a) ranking paradoxes, and (b) requiring extra theoretical machinery to block subsets of hierarchies. The tagged model introduces no new theoretical machinery, and has the added value that it is consistent with psycholinguistic approaches which rely on either suppression or activation. I develop a model which (a) incorporates the syntax/phonology interface and (b) articulates details of the various stages of Phonetic Form (PF) spellout. Such an approach recognizes the important role phonology plays in current models of the grammatical architecture, particularly as applied to multilingual knowledge and performance.

Abstract This chapter explores the L2/L3 acquisition of Mayan ejectives from multiple languages. Proceeding from the dimensional framework of Avery and Idsardi (2001) , we lay out a unified analysis of the phonological acquisition of Yucatec Maya by Spanish speakers, Kaqchikel by Spanish-English speakers, and Q’eqchi’ by English speakers. Both Spanish and English phonologies provide for a path to acquisition of glottalized consonants through their respective contrastive dimensions. Through phonetic completion and enhancement, learners redeploy knowledge of other laryngeal dimensions to successfully acquire sounds which L1 users of the target languages represent with Larynx Height, a dimension not utilized in the learners’ known language(s). Thus, language learners develop non-targetlike grammars that adequately map novel contrasts through the constraints provided by Universal Grammar.

Certain properties of second language (L2) speech are well studied, yet it is uncontroversial to note that L n phonology (where n = any natural number; e.g. L2, L3, etc.) is under-represented in generative approaches to language acquisition compared with the domain of morphosyntax. If, however, we look at L n input and output without taking the learnability of abstract mental representation seriously in our psycholinguistic probes then we miss out on fundamental knowledge as to the nature of a multilingual grammar. L n knowers have complex, phonological grammars whose properties help us to describe and explain their knowledge and behaviour. Many approaches (e.g. usage-based; exemplar) have assumed that phonology can be learned by ‘noticing’ elements in the input. Such a view ignores Plato’s Problem of the acquisition of knowledge as well as the corollary of Orwell’s Problem. Phonology is rich, hierarchical, recursive, governed by UG (universal grammar), and subject to poverty-of-the-stimulus effects. Assigning phonetic tokens to phonological categories entails an algebraic function in which the phonological categories act as variables. Interestingly, this is related to the question of whether phonology is ‘merely’ a system of externalization (which implies it evolved after Merge) or whether there is evidence of it emerging earlier in the lineage of Homo sapiens . I present some arguments that human phonology is not just the linearization of syntax implemented by computationally-simpler, evolutionarily-older machinery. I discuss empirical data which demonstrate the utility of explaining multilingual phonological grammars with reference to hierarchical constituents at the levels of feature, syllable, foot, prosodic word, and phonological phrase; none of these structures are read off the input in a straightforward way. By recognizing the epistemological, representational, and learnability issues related to phonological knowledge (and its interfaces), we deepen our understanding of the full range of the cognitive architecture of the multilingual language faculty.

While the properties of bilingual codeswitching are well-documented, and switching morphology within words is often addressed, the properties of intraword phonology are less-understood. Morphemes from more than one language occur within a single word (e.g., a root from one language and affixes from the other). Furthermore, the affixes come only from the language which generates the syntactic tree while the root can be taken from either language. The data also strongly suggest that within such a morphologically mixed word, the phonology does not switch, a property I refer to as phonological uniformity. The key question explored in this paper is: why is phonological mixing in a morphologically mixed word not allowed? I first present evidence that there is phonological activation of both languages even in a monolingual task such as a Lexical Decision Task or silent reading which is consistent with an integrated I-phonological grammar. I provide a reanalysis of some fascinating data from Delgado et al. (2022) in arguing against a phase-based account of phonological uniformity. The mechanism which ensures that the phonology of the X0 matches the language of the affixes is Match Theory (Selkirk, 2011). The preferential mapping is between (a) syntactic phrases (XPs) and phonological phrases (f), and (b) syntactic heads (X0s) and prosodic words (w). Match Theory’s (monolingual) assumption that syntactic and phonological structure are isomorphic can easily be extended to bilinguals through language tags (Green & Abutalebi, 2013). I recast the Match Theory machinery within the framework of Co-phonologies (Sande, Jenks & Inkelas, 2020). In order to account for the differential behaviour of determiners in codeswitched simple DPs (e.g. the mesa) versus codeswitched complex DPs (e.g. the brown mesa) I show how a combination of the notion of the head of the phonological phrase, the free (as opposed to affixal) clitic status of the English determiner, and the parsing of an English vocabulary item via the Spanish contrastive hierarchy explains the phonological properties observed. A Null Theory phonological account of phonological uniformity is argued to be preferred over a phase-based account.

In this paper, I provide a general response to the points made in the commentaries. One of the major questions probed is whether gradient, variable output is diagnostic of gradient mental representations. I argue that this is not necessarily the case, and look to aspects of the learning theory (such as input processing, restructuring, and cue reweighting), as well as the architecture of the phonology/phonetics interface to account for such variation. I argue for a conservative, incremental restructuring process as the basis of the transition theory of Ln developmental paths. I reiterate the nature of the projection problem at various levels of the prosodic hierarchy when it comes to the input underdetermining the cues to abstract, algebraic phonological constituent labels. The question of whether Ln grammars are consistent with the structural properties of natural languages (i.e. constrained by UG) is discussed. I conclude with a presentation of the idea that linguistic representations can be considered wavelike superstates. This has the potential of capturing what has been described as the fuzziness of representations, as well as the benefit of unifying our treatment of mental and physical objects.

This paper explores the range of contexts in which English is taught as a third language globally. We consider situations where the L3 classroom makeup would be quite fixed (such as a minority L1 and a majority L2; or a majority L1 and a colonial L2) as well as contexts where the L3 classrooms can be quite mixed (such as cases where there is an immigrant population with diverse language histories). Three prominent models of third language acquisition are discussed: the Typological Primacy Model, the L2 Status Factor Model, and the Linguistic Proximity Model.

The purpose of the study was to examine the efficacy of lexical stress diacritics on the English comprehensibility and accentedness of Korean speakers. To this end, 30 native Korean participants read aloud 15 English sentences without diacritics in the pretest. Then, they were given explicit instructions on the production of higher pitch and extended duration as a marker of English stress with musical notation provided. In the posttest, the participants read aloud the same sentences as were in the pretest but which had diacritics indicating stress placement. In the delayed posttest, two days after the pretest and the posttest, the participants read 15 sentences without diacritics again to see if the effects of the treatment were retained. Randomized speech samples were rated by three native speakers of English in relation to comprehensibility and accentedness. Findings showed that significant improvements were observed after the treatment in both comprehensibility and accentedness.

Abstract This chapter explores the construct of a phonological grammar with specific reference to syllable structure and stress. The background literature in the L2 acquisition of sub-syllabic constituents (e.g. onset and coda) is summarized, with illustrations of how differing repair strategies can reveal the influence of markedness and universals. Through the construct of ultimate attainment, the idea of full transfer / full access in phonology is explored. The chapter presents a critique of the influential stress ‘deafness’ literature and its oversimplification of the phonology of stress. The discussion of phonological licensing reveals how the interface of segments and syllables is important in understanding phonological parsing.

Drawing on Archibald (2022a, b), the chapter shows how a contrastive hierarchy model of segmental phonology can provide formal model of determining cross-linguistic similarity. Looking primarily at Arabic-French learners of English, the L1 and L2 features (including the markedness value of the feature) and the rankings influence L3 acquisition. Neurolinguistic and sociolinguistic evidence for the differential behaviour of marked versus unmarked values are discussed, and then it is shown how this variation can act as a cue for the learner to discover the L3 grammatical hierarchy. The author explores a theory of L3 restructuring based on principles of merger, redeployment, and triggering. Ultimately, it is argued that the learner compares the L1 and L2 contrastive hierarchy parses of the L3 input and chooses the one which is optimal for the L3 grammar.

Abstract The study of Lx phonology is under-represented in the GenSLA literature. I argue that this is largely because the complex learnability issues in the phonological domain have been under-appreciated. Drawing on examples from features and syllables, I illustrate that the acquisition of phonology is subject to poverty of the stimulus effects. Consequently, it follows that phonological representations are learned and not merely noticed in the environment. Furthermore, phonological grammars are shown to be generative, and hierarchical. By adopting abstract representational models we can both describe and explain learner behaviour. I explore the role of phonology in morphological and syntactic interfaces, and conclude with a discussion of real-world implications with regard to pronunciation training, and broader societal concerns.

Abstract This chapter demonstrates the benefit of abstract L2 phonological representations by raising problems with solely phonetic models of L2 speech. In articulating an acquisition model, the intake frequency hypothesis is presented which accounts for acquisition by phonetics. The second learning mechanism proposed is the redeployment hypothesis which accounts for acquisition by phonology. The chapter concludes with a discussion of illusory vowels in the production and perception of L2 speech. It is argued that (a) abstract phonological representations (in particular appendices) explain the cross-linguistic patterns, and (b) that inaccurate production is not always caused by inaccurate perception (contra the Speech Learning Model).

Subject Phonetics and Phonology Psycholinguistics Collection: Oxford Scholarship Online

Drawing on Archibald (2022a, b), the chapter shows how a contrastive hierarchy model of segmental phonology can provide formal model of determining cross-linguistic similarity. Looking primarily at Arabic-French learners of English, the L1 and L2 features (including the markedness value of the feature) and the rankings influence L3 acquisition. Neurolinguistic and sociolinguistic evidence for the differential behaviour of marked versus unmarked values are discussed, and then it is shown how this variation can act as a cue for the learner to discover the L3 grammatical hierarchy. The author explores a theory of L3 restructuring based on principles of merger, redeployment, and triggering. Ultimately, it is argued that the learner compares the L1 and L2 contrastive hierarchy parses of the L3 input and chooses the one which is optimal for the L3 grammar.

Abstract This chapter explores the phonology/syntax interface to reveal how phonological matters are critical to understanding aspects of syntactic behaviour in bilinguals. The chapter presents two case studies. The first looks at the role of prosody in sentence processing in hearing and hard-of-hearing populations by focussing on aspects of relative clause attachment preferences and sentential reanalysis. The second looks at the acquisition of L2 WH-questions in learners of Japanese within the framework of Contiguity Theory. The chapter argues that L2 learners construct grammars which are governed by universal properties such as phonological contiguity even though this is not part of the L1 grammar of WH-movement. The resulting grammars are hierarchical and recursive. The chapter demonstrates that the learners acquire the abstract phonology of the L2, though not nativelike phonetic implementation.

Abstract This chapter explores the phonology/morphology interface to reveal how phonological matters are critical to understanding aspects of morphological behaviour in bilinguals. The chapter begins with an exploration of why certain L2 morphemes are omitted in production via the Prosodic Transfer Hypothesis. The framework of Distributed Morphology is introduced to explain aspects of the bilingual lexicon. Traditional psycholinguistic experimental results are presented and then discussed within the context of Distributed Morphology: a case study of L2 Swedish data and a case study of L2 German allomorphy. These case studies show the influence of Universal Grammar on the representations selected by the learners. The chapter concludes with a discussion of intraword codeswitching within a Distributed Morphology account and what these data reveal about the architecture of the multilingual grammar.

In this research note I want to address some misunderstandings about the construct of redeployment and suggest that we need to fit these behavioural data from Yang, Chen and Xiao (YCX) into a broader context. I will suggest that these authors’ work is not just about the failure of three models to predict equivalence classification. Equivalence classification is not the end of the story but only the beginning. We need to look at what cues are detected in the input, which subset of the input becomes intake, and how this intake is parsed onto phonological structures. The empirical results of YCX should not be viewed as some sort of non-result inasmuch as none of the proposed predictors of Mandarin equivalence classification foresaw that the Russian prevoiced stops and short-lag stops would be equated with the Mandarin short-lag stops. Rather, the empirical results need to be contextualized by considering such factors as cue reweighting as part of the learning theory which maps intake onto phonological representations. In this light, the results are not a repudiation of phonological redeployment, but help to shed light on the parsing of the acoustic signal, the importance of robust burst-release cues, and the non-local nature of L2 phonological learning (as opposed to noticing).

Schwartz and Sprouse (2021) argue against property-by-property Transfer (Westergaard, 2021a, b) and for wholesale transfer (Rothman, 2015) into a third language grammar by questioning the cognitive plausibility of "extracting a proper subpart from the horizontal ellipsis grammar and using that proper sub-system as the basis for a new cognitive state." I will argue that the insights from the approaches of Lopez (2020); Lightfoot (2020); Dresher (2018), and Westergaard (2021a) when applied to empirical data from L-3 English data from L-1 Arabic/L-2 French speakers, give us reason to question Schwartz and Sprouse's defence of wholesale transfer, and its typological underpinnings. We can set the study of L3(A) in a larger context which can unify domains such as the acquisition of phonology and syntax via a unified approach to parsing. By invoking an underspecified, minimal UG, primary linguistic data, and domain-general third factors which act in concert to parse the E-language to select structures, we can capture the underlying similarity of first, second, and third language acquisition. Parsing proceeds in an error-driven fashion, structure by structure, drawing on the Integrated I-language and UG options found in a Repository. In essence, this approach renders the wholesale/property-by-property distinction a false dichotomy.

Abstract The main goal of this book is to probe questions about the nature of an interlanguage (IL) grammar (i.e. the grammar of a bilingual or multilingual). I approach these questions from a cognitive science perspective which draws upon abstract representational structures in demonstrating that phonological knowledge underlies the surface phonetic properties of L2 speech. Specifically, the book will demonstrate that IL grammars are not ‘impaired’, ‘fundamentally different’, or ‘shallow’ (as some have argued). The phonological grammars are complex, hierarchically structured mental representations that are governed by the principles of linguistic theory, including the principles of Universal Grammar. I craft a model which addresses Plato’s Problem (learning in the absence of evidence) and Orwell’s Problem (resistance to learning in the face of abundant evidence). Furthermore, the study of grammatical interfaces (phonetics/phonology; phonology/morphology; phonology/syntax) reveals the necessary design conditions for an internally consistent architecture for a comprehensive model of second language speech. The resulting empirically motivated model is parsimonious in accounting for all aspects of L2 speech from phonological feature, to segment, to word, to sentence. The book concludes with discussion of why phonology has been underrepresented in generative approaches to second language acquisition, as well as some of the implications of second language phonology for applied linguistics and language pedagogy.

In this article, I tackle the question of differential substitution in L2 phonology. A classic example of the phenomenon is learners from different L1s attempting to acquire the L2 English interdental fricative /θ/. Speakers of some languages (e.g., Japanese) tend to pronounce the /θ/ as [s] while speakers of other languages (e.g., Russian) tend to pronounce the /θ/ as [t]). Since both Japanese and Russian have both /s/ and /t/ in their phonemic inventories, it is interesting to ask why one language would choose [s] and the other [t]. What I argue in this article is that it is not a local comparison of two sounds, two features, or two phonemes that will determine why one segment rather than another is substituted. Rather, I argue that we must consider the formal representation of the entire segmental inventory (represented as a contrastive hierarchy) in order to understand why the Japanese pick the [s] but the Russian the [t] as the “best” substitute for the English /θ/. What I will demonstrate is that in the languages that substitute [s], [continuant] is the highest-ranked feature that has scope over the place and voice features in the contrastive hierarchy of phonological features. In the languages that substitute [t], the place and voice features rank above [continuant].