Non-Dravidian Elements and (Non)Diasystematic Change in Malayalam
Total Page:16
File Type:pdf, Size:1020Kb
To appear in Constructions in Contact 2: Language change, multilingual practices, and additional language acquisition eds. Höder, S. & Boas, H. Title: Non-Dravidian elements and (non)diasystematic change in Malayalam Savithry Namboodiripad (University of Michigan, Ann Arbor) [email protected] Abstract: This chapter applies a Diasystematic Construction Grammar (DCxG) approach to account for non-Dravidian vocabulary and phonology in Malayalam, a high-contact Dravidian language. The distinction made in DCxG between diaconstructions, which are language non-specific, and idioconstructions, which are language-specific, proves useful in accounting for semantic specialization and phonological heterogeneity due to language contact. Notably, increased contact with English has led in some cases to decreased phonological adaptation, as some constructions change from diaconstructions to idioconstructions: Non-diasystematic change. Taken together, this chapter argues that any analysis of Malayalam must account for non-Dravidian subpatterns, and including language labels as part of speakers' linguistic knowledge enhances our understanding of the dynamics of language contact. Keywords: Malayalam, English, language contact, language change, semantic specialization, loanword adaptation, historical linguistics, sociolinguistics 1 High-contact languages are not categorically different By introducing a distinction between diaconstructions, which are language non-specific, and idioconstructions, which are language-specific, Diasystematic Construction Grammar (DCxG) captures the fact that borders between languages, while often blurred, also shape language use (Höder 2012). Building on other construction-theoretic approaches, DCxG characterizes all languages as consisting of emergent subpatterns of constructions. This includes hybrid or high-contact languages, which differ only in that those subpatterns can be attributed (by analysts, not necessarily speakers) as belonging to a particular language. As such, DCxG provides an ideal framework for analyzing high-contact languages while also capturing smaller-scale contact effects. Crucially for those studying language contact, this obviates the need for a fundamental distinction between high- and low-contact languages (e.g., Faraclas & Klein 2009, Mufwene 2000). 1 To appear in Constructions in Contact 2: Language change, multilingual practices, and additional language acquisition eds. Höder, S. & Boas, H. In this chapter, I apply a DCxG approach to account for non-Dravidian vocabulary and phonology in Malayalam, a Dravidian language with considerable influence from Sanskrit and English, as spoken in Kerala, India.1 The outline of the chapter is as follows: The remainder of Section 1 discusses the data and goals for the chapter. Section 2 covers the sociohistorical contexts of the language contact which led to the non- Dravidian elements in Malayalam, culminating in contemporary attitudes towards English as documented in sociolinguistic interviews. Sections 3 and 4 discuss how accounting for word origin contributes to our understanding of semantic change and phonological heterogeneity respectively in Malayalam. In Section 5, I show how increased language contact can lead to non-diasystematic change via loanword de-adaptation. Section 6 focuses on word-final schwa as a signifier of “Malayalam” idioconstructions. Finally, Section 7 advocates for Malayalam as a hybrid language. 1.1 “Stratified lexicon” without a lexicon As demonstrated in Section 3, Sanskrit- and English-origin words in Malayalam are highly frequent and appear across a variety of semantic domains. Analogous to Latinate and Germanic morphophonological subpatterns in English, Sanskrit and English words in Malayalam have phonological patterns which are not found in Dravidian-origin words, such as heteroorganic clusters, certain codas, and voiced aspirated stops. Non-constructional approaches have dealt with this heterogeneity due to language contact in Malayalam in a variety of ways. Previous analyses of Malayalam phonology and morphology (Mohanan, T., 1989; Mohanan & Mohanan 1986) have posited distinct derivational processes for Dravidian- associated and Sanskrit-associated words, dependent on whether the words have a [+DRAVIDIAN] or [+SANSKRIT] feature. This approach anticipates a main insight of DCxG, in that a speaker’s analysis of a word being of Sanskrit or Dravidian origin has an effect on the constructions it is connected to (or, in the derivational approach, in the rules and rule orderings which apply). Crucially, speakers’ analyses may or may not align with a word’s etymological origin as determined by linguists; speakers pay attention to a variety of phonological, meaning-based, and social cues in their assessment of a word. 1 Malayalam (Glottocode: mala1464) is a South Dravidian language spoken by about 35 million speakers in India (where it has official designation, and the status of a Classical Language), Lakshadweep, Puducherry, and by a considerable diasporic population all over the world. 2 To appear in Constructions in Contact 2: Language change, multilingual practices, and additional language acquisition eds. Höder, S. & Boas, H. In constraint-based approaches, language-specific patterns are abandoned in favor of a core-periphery distinction, in which indexed constraints are differentially violated in distinct strata of the lexicon (Itô & Mester 1993). From this perspective, the perceived etymology of words is not part of speakers’ representations, rather, the strata are emergent from the constraints which are differentially violated. Nativization is defined as the violation of fewer constraints, as words move from the periphery -- in which more constraints are violated -- to the core, where fewer constraints are violated. Itô & Mester specifically refer to the Sanskrit words in Malayalam as being an instantiation of core-periphery stratification in the lexicon. In such approaches, constraints are theoretical primitives; therefore, patterns in the lexicon emerge from the interaction between constraints. In constructional approaches, patterns emerge from inferences across learned form-meaning mappings, constructions, which are richly specified with information that speakers use to make comparisons and connections. The crucial analytic contribution of DCxG is that speakers’ analyses of what language a particular construction belongs to can be part of the construction itself. This approach assumes rich lexical knowledge, and, following Elman’s model of lexical knowledge without a lexicon, the enterprise becomes psychologically tractable (Elman 2009). Social information such as language origin can be part of a speaker’s lexical knowledge (e.g., Kristiansen 2008), even without a stored lexical entry. When describing the selective violation of “native” phonological patterns, as found in Malayalam, a DCxG approach allows for an analytic move from the idea of a stratified lexicon toward a model which contains networks of connected constructions which can or cannot be associated with a particular language. Diasystematic change, in which language-specific constructions, idioconstructions, become language-non-specific diaconstructions is one possible outcome of language contact: Elements which were previously marked in some way (whether in their form or function) as being associated with a particular language adapt to the larger constructional networks in the language. This is one way of describing what happens during loanword adaptation, for example. While we will see such examples in this chapter, we will also see cases of language contact resulting in non-diasystematic change (Section 5), that is, previously language-non-specific constructions becoming language-specific. Taken together, this chapter argues not only that any analysis of Malayalam must account for non-Dravidian subpatterns, but that accounting for perceived etymology and/or word origin using a DCxG approach enhances our understanding of the dynamics of language contact. 3 To appear in Constructions in Contact 2: Language change, multilingual practices, and additional language acquisition eds. Höder, S. & Boas, H. 1.2 Note on the data There is not yet a tagged corpus of spoken Malayalam; written standard Malayalam differs considerably from spoken Malayalam, and, as described in the following section, affinity towards and use of written Malayalam is in flux. Online (on WhatsApp, Facebook, and other social media platforms), written standard Malayalam is used in addition to transliteration of Malayalam in English script, and transliteration of English and spoken Malayalam in Malayalam script. Unless cited otherwise, the data in this chapter comes from observations of Malayalam usage online, via WhatsApp, in popular culture, and from field work conducted in 2014 and 2016 in central Kerala – mostly in areas surrounding Thrissur (Trichur)2 and Kochi (Cochin). Malayalam words are represented using conventional transliterations, or, to avoid confusion due to variance in transliteration conventions, using the International Phonetic Alphabet. All transcriptions are broad and my own. Because valid measures of frequency are not available for spoken varieties of Malayalam, no strong claims about relative frequency or the effect thereof are made here. However, care was taken to include words which are in common use. A word was determined to be ‘no longer in common use’ if it was not found in internet searches of magazines, newspapers, blogs and other Malayalam media online,