Joseph Brooks
My fieldwork is based among the Awakngi people in Andamang village on the Sogeram, in inland Madang in PNG, where I've been studying their language, Chini, since 2012. In my research I am motivated to understand language and culture together and so rely upon a diverse toolkit of methods and theoretical insights from both linguistics and anthropology. The areas I consider my work to be in closest conversation with are: anthropological linguistics, functional linguistics, language documentation, linguistic anthropology, and linguistic description.
Supervisors: Marianne Mithun, Bernard Comrie, Wallace Chafe, and Marian Klamer (former Ph.D. advisors) Lise Dobrin (former post-doctoral advisor)
Supervisors: Marianne Mithun, Bernard Comrie, Wallace Chafe, and Marian Klamer (former Ph.D. advisors) Lise Dobrin (former post-doctoral advisor)
less
Related Authors
InterestsView All (15)
Uploads
Papers by Joseph Brooks
This sketch grammar is based on three months of fieldwork in Iteri, a Left May language with approximately 500 speakers spoken near the East Sepik-Sandaun border in a northern spurt of the highlands called the West Range. Examples are from my fieldnotes from participant observation and elicitation as well as 1.5 hours of annotated narrative and conversational speech.
Since the days of the Boasian trilogy over a century ago, transcription has been acknowledged as central to field linguistics. Field linguists have long recognized that without it, it is all but impossible to understand the workings of certain grammatical constructions (Himmelmann 1998; Mithun 2014, 2001), something also emphasized in my own research (Brooks 2018). Although transcription has emerged as the core research activity in documentary linguistics and constitutes the major workload (Himmelmann 2008, 2018; Jung & Himmelmann 2011), the tedious, laborious nature of transcribing connected speech means that scientific ideals of maximal annotation of documentary corpora are far from realistic. Thus far, resolution has been located in computational methods which might (or might not) streamline the documentation process (Bird 2020; Cox 2019; Brooks 2013; Palmer et al. 2010). Meanwhile, based on her linguistic anthropological fieldwork among the Bosavi, Schieffelin (1990) comes to an altogether different conclusion about transcription, arguing it is best understood as an ethnographic process. Here I rely on my research and relationships with Chini people of Andamang village (inland Madang, PNG) to show how my own methodological preoccupation with the transcription of connected speech – that is, the result of my own research goals and the output demands of my funding agency – had unintended social consequences which I was late to identify and rectify.
I begin with an overview of the theory and practice of transcription in linguistics and anthropology, drawing attention to several variables concerning transcription in fieldwork. I then discuss how the role of transcription in my own research, though it produced one of the largest collections of annotated conversational data for a New Guinea language, unfortunately by-produced a socially-methodologically fraught form of fieldwork. I address three interrelated (social-methodological) problems: (1) disproportionate exchange relations, effects on local forms of labor, and feud exacerbation; (2) the introduction of distortions for traditional Chini conceptions of language; and (3) the insufficiency of extensive transcribed speech in the linguistic analysis of particular constructions in Chini. I describe the ways in which these problems exposed a conflict between native Awakng’i (i.e., Andamang Chini) hopes and desires about my participation in local social life vis-a-vis the scientific perspective I was immersed within which sees little value in the social immersion and interpretative task of the fieldworker, instead supporting what Dobrin & Schwartz (2016) identify as the ‘objectivity paradigm’. I explain how these problems have come to find their resolution in participatory methods similar to the discussion found in the likes of Nida (1957), Everett (2001), and Sarvasy (2015). I conclude with some discussion points about the role of transcription vis-a-vis participation.
This paper is structured as follows.
1. Pragmatic functions versus switch-reference
After giving an overview of the clause chaining constructions, I begin the discussion by addressing switch reference, which is of course the traditional view of the basis of Papuan-style clause chaining systems. I explain how switch reference as a possible analysis for the continuous information (=kɨ & =tɨ) and presuppositional asymmetry (=va & =mɨ), fails to explain the many exceptions to the expected marking (same-subject and different-subject respectively). Instead, as other scholars have shown for clause linking constructions in other languages, continuous versus discontinuous reference is only part of the etic material in Chini; that is, it is a secondary effect of continuous information on the one hand and pragmatic asymmetry on the other. I discuss how this is especially visible in conversational speech, where a much wider range of combinations of referents occurs than in many genres of narrative.
I also discuss how evidence showing that Chini (unlike any other known Lower-Sepik Ramu language) acquired its clause chaining system through contact with neighboring Sogeram peoples, adds credence to this analysis.
2. Temporal succession and a twist
Here I discuss the linkers of temporal succession =ndaka (realis) & =ndata (irrealis). I also discuss a further division of these, one that is often impossible to retrieve from digital recordings. It turns out there is a further division of both linkers: =ndaka & =ndakɨ and then =ndata & =ndatɨ. I present examples primarily from unrecorded interactions that depended upon my knowledge of Chini and my understanding of the full pragmatic context of the speech event, to learn that speakers rely on this possibility to subtly express information about the perceived likelihood of the successive event.
3. The contribution of realis and irrealis marking
Unlike in any other described Papuan language, realis and irrealis is fully part and parcel of clause chaining; as the linkage pairs themselves demonstrate, the speaker is obliged to choose one or the other for each pair of linkers. After giving a brief account of the language-specific meaning of this distinction, I discuss the types of information realis and irrealis marking contribute in Chini clause chains, the constructional dexterity found in the pragmatic concord they express, and finally, mid-chain realis-irrealis shifts, where the otherwise robust concord rules, are dispensed with.
I also discuss the role of contact and how Chini appears to have extended the realis/irrealis marking limited in Sogeram languages to different-subject medials (in some Sogeram languages, with further limitations). This extension of a realis/irrealis distinction, one already salient elsewhere in Chini grammar, resulted in a novel system where realis/irrealis is distinguished throughout as a fundamental component of all chaining constructions in the language.
4. Medial and final clauses in terms of dependency
Based on 1-3, here I discuss the ways in which Chini medial and final clauses conform to notions of dependency in terms of their semantic, pragmatic, and syntactic properties. In this section I also discuss two further construction types, namely the "stacking" construction where =kɨ (continuous information, realis) & =va (pragmatic asymmetry, realis) co-occur forming =vakɨ, and then the use of =vakɨ, =kɨ, and =mɨ medial constructions as independent clauses.
5. Conclusion
I situate the above discussion about Chini in the context of what is known about Papuan clause chaining, discussing how the pragmatic basis of the constructions, the realis/irrealis component, and other aspects of Chini dependency relations, represent significant expansions of our knowledge of what speakers of Papuan languages use clause chaining constructions to do.
One of these is that the soil of the lower Sogeram, radically unlike the much larger Ramu river which it empties into, benefits from the flow of nutrients during the wet season (November-March) from the entire easten stretch of the Sogeram. This means that residents of the Sogeram have much richer soil in which a much greater diversity of crops can be grown than is the case for the (Rao-speaking) people whose villagers are upriver (south) on the Ramu, from the confluence of the Sogeram and Ramu rivers. The second geological benefit of the local Sogeram is the presence of abundant oxbow marshes. These marshes, most of which are located in Andamang (Chini-speaking) and to some extent Limbebu (Breri-speaking) are created by the changes that occur from one wet season to the next, as the meander belt of the lower Sogeram shifts and redirects its path, on occasion dramatically so. Old chunks of the river become severed from the new meander belt, all fish and other wildlife (e.g. crocodiles) are then preserved, providing the local people with an easy source of protein.
Here I discuss how the unmatched geological wealth of Andamang territory squares with linguistic evidence and evidence from local history, which show how this well sought after chunk of land has been a small but significant nexus of contact and settlement by multiple Papuan groups over the past centuries.
Taking into account the breadth of the literature on the topic including grammars of languages described as having realis-irrealis distinctions and/or realis or irrealis as lone categories, I focus the discussion on Chini spoken in inland Madang, Papua New Guinea, a language in which realis-irrealis distinctions are unusually elaborated, perhaps moreso than any other described language. This is because in Chini the distinction is marked not only in multiple parts of the verb morphology but also separately in the clause chain linkage enclitics.
I show: (1) that temporal reference is not part of the emic meaning in realis-irrealis distinctions in Chini; (2) that there is an important relation between the area of the grammar where the distinction occurs and its particular division of the semantic-pragmatic space; and (3) that realis-irrealis distinctions are not quite identical to (and thus not necessarily comparable with) phenomena covered by 'realis', 'irrealis' and other similar labels whenever these involve (lone) categories. This is because they lack the critical dualism that characterizes the distinction, as it occurs in Chini and in the grammars of many other languages.
In addition to these findings I present an overview of the history of scholarship on this topic over the last century, discuss various other components of the Chini workings and issues of cross-linguistic comparability, and describe relevant parts of Chini grammar and aspects of Chini culture.
Talks by Joseph Brooks
The current literature on clause chaining in New Guinea languages holds that for languages which distinguish realis and irrealis clause chaining linkers, the choice of one or the other type on previous medial clauses, is determined by agreement with verbal categories in the final clause. In this paper I rely on documentary linguistic evidence from Chini to support a different analysis, namely that speakers produce the linkers online in a way that can preclude governance from information in an upcoming final clause. Although most chains do contain only realis or irrealis linkers, Chini speakers nonetheless do freely shift back and forth between realis and irrealis linkers within single chains.
I also suggest why different types of linguistic data might be at play in leading different linguists to different analyses of agreement phenomena in Papuan-style clause chaining constructions. Whereas the agreement analysis for realis and irrealis marking has been based on elicited and narrative data, the evidence I have used comes entirely from naturalistic Chini conversation, which lacks a 'script' and often involves much more complex interrelations between events and their expression in clause chaining constructions. Although the match between information in final clauses and realis vs irrealis linkers on previous medial clauses does often permit analysis in terms of concord, conversational data from these types of languages shows that the workings of these constructions is more complex than previously understood.
Whereas much of linguistics and language documentation have come to value annotated digital corpora as the ideal empirical basis for linguistic claims, I discuss how methods which are under-theorized and even devalued in linguistics were key in my analysis of two sequential clause chain linkage constructions in Chini, a Lower Sepik-Ramu language of inland Madang, PNG.
The discussion centers around a pragmatic minimal pair of two biclausal chains from two unrecorded spontaneous Chini conversations during my 2018-2019 fieldwork, conversations in which I was myself a participant. These examples do not really represent "empirically sound" data; they are not valuable from the perspective of resolvability or objectivity as scientific ideals. Instead, the intellectual "low ground" of social immersion and language learning turned out to be invaluable in discovering the subtle but key linguistic difference between these clause chaining constructions, a difference which, though present among the ~800 tokens in the Chini corpus, cannot be readily accounted for by native speakers who are not linguists nor determined from the digital annotations (by a linguist who is not a native speaker). I explain why the documentary data have an important ancillary role in supporting analyses that draw centrally from social immersion, but can in fact be quite limited in their potential to give insight into language-specific pragmatic workings such as those found in the Chini linkage constructions.
This sketch grammar is based on three months of fieldwork in Iteri, a Left May language with approximately 500 speakers spoken near the East Sepik-Sandaun border in a northern spurt of the highlands called the West Range. Examples are from my fieldnotes from participant observation and elicitation as well as 1.5 hours of annotated narrative and conversational speech.
Since the days of the Boasian trilogy over a century ago, transcription has been acknowledged as central to field linguistics. Field linguists have long recognized that without it, it is all but impossible to understand the workings of certain grammatical constructions (Himmelmann 1998; Mithun 2014, 2001), something also emphasized in my own research (Brooks 2018). Although transcription has emerged as the core research activity in documentary linguistics and constitutes the major workload (Himmelmann 2008, 2018; Jung & Himmelmann 2011), the tedious, laborious nature of transcribing connected speech means that scientific ideals of maximal annotation of documentary corpora are far from realistic. Thus far, resolution has been located in computational methods which might (or might not) streamline the documentation process (Bird 2020; Cox 2019; Brooks 2013; Palmer et al. 2010). Meanwhile, based on her linguistic anthropological fieldwork among the Bosavi, Schieffelin (1990) comes to an altogether different conclusion about transcription, arguing it is best understood as an ethnographic process. Here I rely on my research and relationships with Chini people of Andamang village (inland Madang, PNG) to show how my own methodological preoccupation with the transcription of connected speech – that is, the result of my own research goals and the output demands of my funding agency – had unintended social consequences which I was late to identify and rectify.
I begin with an overview of the theory and practice of transcription in linguistics and anthropology, drawing attention to several variables concerning transcription in fieldwork. I then discuss how the role of transcription in my own research, though it produced one of the largest collections of annotated conversational data for a New Guinea language, unfortunately by-produced a socially-methodologically fraught form of fieldwork. I address three interrelated (social-methodological) problems: (1) disproportionate exchange relations, effects on local forms of labor, and feud exacerbation; (2) the introduction of distortions for traditional Chini conceptions of language; and (3) the insufficiency of extensive transcribed speech in the linguistic analysis of particular constructions in Chini. I describe the ways in which these problems exposed a conflict between native Awakng’i (i.e., Andamang Chini) hopes and desires about my participation in local social life vis-a-vis the scientific perspective I was immersed within which sees little value in the social immersion and interpretative task of the fieldworker, instead supporting what Dobrin & Schwartz (2016) identify as the ‘objectivity paradigm’. I explain how these problems have come to find their resolution in participatory methods similar to the discussion found in the likes of Nida (1957), Everett (2001), and Sarvasy (2015). I conclude with some discussion points about the role of transcription vis-a-vis participation.
This paper is structured as follows.
1. Pragmatic functions versus switch-reference
After giving an overview of the clause chaining constructions, I begin the discussion by addressing switch reference, which is of course the traditional view of the basis of Papuan-style clause chaining systems. I explain how switch reference as a possible analysis for the continuous information (=kɨ & =tɨ) and presuppositional asymmetry (=va & =mɨ), fails to explain the many exceptions to the expected marking (same-subject and different-subject respectively). Instead, as other scholars have shown for clause linking constructions in other languages, continuous versus discontinuous reference is only part of the etic material in Chini; that is, it is a secondary effect of continuous information on the one hand and pragmatic asymmetry on the other. I discuss how this is especially visible in conversational speech, where a much wider range of combinations of referents occurs than in many genres of narrative.
I also discuss how evidence showing that Chini (unlike any other known Lower-Sepik Ramu language) acquired its clause chaining system through contact with neighboring Sogeram peoples, adds credence to this analysis.
2. Temporal succession and a twist
Here I discuss the linkers of temporal succession =ndaka (realis) & =ndata (irrealis). I also discuss a further division of these, one that is often impossible to retrieve from digital recordings. It turns out there is a further division of both linkers: =ndaka & =ndakɨ and then =ndata & =ndatɨ. I present examples primarily from unrecorded interactions that depended upon my knowledge of Chini and my understanding of the full pragmatic context of the speech event, to learn that speakers rely on this possibility to subtly express information about the perceived likelihood of the successive event.
3. The contribution of realis and irrealis marking
Unlike in any other described Papuan language, realis and irrealis is fully part and parcel of clause chaining; as the linkage pairs themselves demonstrate, the speaker is obliged to choose one or the other for each pair of linkers. After giving a brief account of the language-specific meaning of this distinction, I discuss the types of information realis and irrealis marking contribute in Chini clause chains, the constructional dexterity found in the pragmatic concord they express, and finally, mid-chain realis-irrealis shifts, where the otherwise robust concord rules, are dispensed with.
I also discuss the role of contact and how Chini appears to have extended the realis/irrealis marking limited in Sogeram languages to different-subject medials (in some Sogeram languages, with further limitations). This extension of a realis/irrealis distinction, one already salient elsewhere in Chini grammar, resulted in a novel system where realis/irrealis is distinguished throughout as a fundamental component of all chaining constructions in the language.
4. Medial and final clauses in terms of dependency
Based on 1-3, here I discuss the ways in which Chini medial and final clauses conform to notions of dependency in terms of their semantic, pragmatic, and syntactic properties. In this section I also discuss two further construction types, namely the "stacking" construction where =kɨ (continuous information, realis) & =va (pragmatic asymmetry, realis) co-occur forming =vakɨ, and then the use of =vakɨ, =kɨ, and =mɨ medial constructions as independent clauses.
5. Conclusion
I situate the above discussion about Chini in the context of what is known about Papuan clause chaining, discussing how the pragmatic basis of the constructions, the realis/irrealis component, and other aspects of Chini dependency relations, represent significant expansions of our knowledge of what speakers of Papuan languages use clause chaining constructions to do.
One of these is that the soil of the lower Sogeram, radically unlike the much larger Ramu river which it empties into, benefits from the flow of nutrients during the wet season (November-March) from the entire easten stretch of the Sogeram. This means that residents of the Sogeram have much richer soil in which a much greater diversity of crops can be grown than is the case for the (Rao-speaking) people whose villagers are upriver (south) on the Ramu, from the confluence of the Sogeram and Ramu rivers. The second geological benefit of the local Sogeram is the presence of abundant oxbow marshes. These marshes, most of which are located in Andamang (Chini-speaking) and to some extent Limbebu (Breri-speaking) are created by the changes that occur from one wet season to the next, as the meander belt of the lower Sogeram shifts and redirects its path, on occasion dramatically so. Old chunks of the river become severed from the new meander belt, all fish and other wildlife (e.g. crocodiles) are then preserved, providing the local people with an easy source of protein.
Here I discuss how the unmatched geological wealth of Andamang territory squares with linguistic evidence and evidence from local history, which show how this well sought after chunk of land has been a small but significant nexus of contact and settlement by multiple Papuan groups over the past centuries.
Taking into account the breadth of the literature on the topic including grammars of languages described as having realis-irrealis distinctions and/or realis or irrealis as lone categories, I focus the discussion on Chini spoken in inland Madang, Papua New Guinea, a language in which realis-irrealis distinctions are unusually elaborated, perhaps moreso than any other described language. This is because in Chini the distinction is marked not only in multiple parts of the verb morphology but also separately in the clause chain linkage enclitics.
I show: (1) that temporal reference is not part of the emic meaning in realis-irrealis distinctions in Chini; (2) that there is an important relation between the area of the grammar where the distinction occurs and its particular division of the semantic-pragmatic space; and (3) that realis-irrealis distinctions are not quite identical to (and thus not necessarily comparable with) phenomena covered by 'realis', 'irrealis' and other similar labels whenever these involve (lone) categories. This is because they lack the critical dualism that characterizes the distinction, as it occurs in Chini and in the grammars of many other languages.
In addition to these findings I present an overview of the history of scholarship on this topic over the last century, discuss various other components of the Chini workings and issues of cross-linguistic comparability, and describe relevant parts of Chini grammar and aspects of Chini culture.
The current literature on clause chaining in New Guinea languages holds that for languages which distinguish realis and irrealis clause chaining linkers, the choice of one or the other type on previous medial clauses, is determined by agreement with verbal categories in the final clause. In this paper I rely on documentary linguistic evidence from Chini to support a different analysis, namely that speakers produce the linkers online in a way that can preclude governance from information in an upcoming final clause. Although most chains do contain only realis or irrealis linkers, Chini speakers nonetheless do freely shift back and forth between realis and irrealis linkers within single chains.
I also suggest why different types of linguistic data might be at play in leading different linguists to different analyses of agreement phenomena in Papuan-style clause chaining constructions. Whereas the agreement analysis for realis and irrealis marking has been based on elicited and narrative data, the evidence I have used comes entirely from naturalistic Chini conversation, which lacks a 'script' and often involves much more complex interrelations between events and their expression in clause chaining constructions. Although the match between information in final clauses and realis vs irrealis linkers on previous medial clauses does often permit analysis in terms of concord, conversational data from these types of languages shows that the workings of these constructions is more complex than previously understood.
Whereas much of linguistics and language documentation have come to value annotated digital corpora as the ideal empirical basis for linguistic claims, I discuss how methods which are under-theorized and even devalued in linguistics were key in my analysis of two sequential clause chain linkage constructions in Chini, a Lower Sepik-Ramu language of inland Madang, PNG.
The discussion centers around a pragmatic minimal pair of two biclausal chains from two unrecorded spontaneous Chini conversations during my 2018-2019 fieldwork, conversations in which I was myself a participant. These examples do not really represent "empirically sound" data; they are not valuable from the perspective of resolvability or objectivity as scientific ideals. Instead, the intellectual "low ground" of social immersion and language learning turned out to be invaluable in discovering the subtle but key linguistic difference between these clause chaining constructions, a difference which, though present among the ~800 tokens in the Chini corpus, cannot be readily accounted for by native speakers who are not linguists nor determined from the digital annotations (by a linguist who is not a native speaker). I explain why the documentary data have an important ancillary role in supporting analyses that draw centrally from social immersion, but can in fact be quite limited in their potential to give insight into language-specific pragmatic workings such as those found in the Chini linkage constructions.