Roberto Zariquiey and Pilar M. Valenzuela: The Grammar of Body-Part Expressions. A View from the Americas

1 Overview

The articles published in this collection, as the editors Roberto Zariquiey and Pilar M. Valenzuela inform in their introductory chapter, are for the most part elaborations of talks delivered at a workshop held at the Pontificia Universidad Católica del Perú in Lima in 2016. Self-avowedly, this volume is not, or not only, on body part semantics, colexification patterns, etc., but rather the less-explored grammar of body parts. The volume presents a rich American perspective on that topic. Reflecting the interests and foci of the editors and the location in which the workshop was held, most chapters in the volume are concerned with South American languages, though Mesoamerica and North American languages also play a role.

2 The individual chapters

The volume is structured topically. It has two parts, dubbed “Categorialization [sic!], lexicalization, and semantic processes associated with body-part expressions” and “The grammar of body-part expressions” respectively.

The opening chapter by Christian Lehmann, however, stands outside these parts. It aims to provide “Foundations of body part grammar” generally, without being tied specifically to the languages of the Americas. The “anthropology” that underlies the chapter, is that, “[A] person has somatic, psychic, and social aspects” (p. 18) and that “[a] body part is, in the first place, a physical object” (p. 15) that is, however, characterized by Lehmann as “non-prototypical” in that it is not clearly individuated. In the discussion of meronymic relations, one again finds the characterization of the body as “a thing” (p. 20). This not only raises the issue of languages that lack a term for “body” (Wilkins 1996) but slights whole research traditions in philosophy and anthropology which suggest that the body is not just a “thing” among others. This objectivizing perspective is not the primary mode in which humans conceive the body and its parts, but a secondary, “unnatural” post hoc classification that takes a significant amount of abstraction.

Cognitively, the domain of body parts, according to Lehmann, is prototypically structured: the most prototypical members are external and under control, and less prototypical ones internal and/or not under control. Lehmann argues that the grammar associated with body-part terms “exhibits certain interlingual regularities” (p. 15), and that where variation exists, it is principled and grounded in aspects of cognition like this. The task Lehmann sets out for himself is to sketch out these regularities and their bases. As an important contribution to the study of body-part grammars cross-linguistically, Lehmann marshals the animacy hierarchy (or, as he prefers to call it, “empathy hierarchy”), otherwise used to explain phenomena like differential object marking, to show how body parts as possessed individual objects rank lower than the possessor, i.e. the person. More generally, the chapter is admirably detailed. In particular the discussion of how the relationship of the whole body and its parts can be coded lexically and morphosyntactically is mapped out in great detail. That being so, it can indeed form a future foundation for the typologization of the behaviour of individual languages. At the same time, the chapter is very bulky, covering 62 pages of the volume and reaching five levels of headings and subheadings, and there is no outline in the introduction that would help the reader to create a mental roadmap of what is to come.

The first main part of the volume opens with Matt Coler, Bertie Kaal, and Edwin Banegas Flores’s analysis of the body-part term nayra ‘eye’ in Muylaq’ Aymara. In a manner that is typical for many languages of the Central Andes, nayra is an “ambivalent”, i.e. it can without further derivation take nominal and verbal morphology and accordingly be interpreted as nominal or verbal. As a noun, nayra translates ‘eye’, but can also refer to the ‘east’, and, usually accompanied by the loanword timpu ‘time’, to the past. Coler et al.’s chapter is predicated on a rich background of work in cognitive psychology on the conception of time that Aymara enshrines. In Aymara, time is divided into the unseen future and the visible and experienced present and past, against the background of a coherent spacetime expressed by the term pacha. This chapter is based on Aymara narratives and weaves together cognitive psychology and linguistic analysis expertly, making, among other things, the point that commonly posited metaphors in the (early) Cognitive Linguistics literature like time as space do not apply to Aymara.

In “Pathways and patterns of metaphor and metonymy in Mixtepec-Mixtec body-part terms”, Jack Bowers discusses extensions of body-part terms to access conceptually abstract semantic domains and the associated grammaticalization processes in an Oto-Manguean language of Mesoamerica, adopting a Cognitive Linguistic and Embodiment perspective. According to Bowers, body-part terms occur frequently in compounds with “a semi-lexicalized sense” (p. 100) to express meronymic relationships between objects and their parts, e.g. titsi-mesa ‘the underside of a table, underneath a table’ (from titsi ‘belly’) and/or come to be used in an adposition-like manner, as when tsa’a ‘foot’ is used to indicate location under something. Of general interest to typologists might be the claim that the development of nuu ‘face’ to function as a generic noun simply meaning ‘place’ represents a case of degrammaticalization (that might be controversial, since, on the semantic side, the item is bleached of virtually all of its original specific meaning as is typical for grammaticalization). The chapter presents rich analyses; it is just a pity that the broader Mesoamerican perspective (many of the processes discussed are common also in other languages in the area, e.g. Levinson 1994) is not established and thus the areal context of the discussed phenomena remains concealed from readers who are not familiar with the literature.

The third chapter, “Body parts in Toba: from the biological to the emotional domain”, by Paola Cúneo and Cristina Messineo, discusses cases in which “figurative expressions in which the condition or the literal behaviour of the body part suggests a condition or figurative behaviour of the person” (p. 126) in this Guaicuruan language. The data reveal few surprises to someone familiar with the literature on figurative language associated with body-part terms (the tongue, for instance, is involved as the anchor in expressions relating to speaking truthfully and deceptively; the heart in expressing emotional or cognitive states), though the chapter is valuable for its detailed, focused descriptive account. The succinct generalization made by the authors is that “visible body parts denote emotional behaviours or attributes related to the social domain and the interaction with others, while the nonvisible body parts involve psychological states or behaviours linked to the individual and more intimate aspects of the person” (p. 135). This statement might be made for other languages as well but is of particular interest for Toba because internal and external body parts also systematically select for different deictic classifiers and thus the difference is not purely semantic, but active in the grammar as well.

In “Body-part categorization and grammar in Piaroa”, Jorge Emilio Rosés Labrada offers a detailed descriptive account of body-part terms and their grammar in this Saliban language, for which he relied primarily on elicitation utilizing the stimuli developed in the 2000s at the Max Planck Institute for Psycholinguistics (Enfield 2006; van Staden and Majid 2006). The data are very rich, and full of interesting details: tuhuoya̧bi̧ ‘knee’ can never be possessed; the language has an apparently monomorphemic term, uhuoro, for ‘seminal vesicle’; in elicitation at least, the quality of mucus can be distinguished between liquid, greasy, and hard via classifiers that are joined to a dummy root. The language shows mappings between animal and human body-part terms, including ‘tail’ – ‘penis’, ‘feathers’ – ‘hair’, and ‘beak’ – ‘mouth’ (see Urban 2012 for the broader South American context). It is a pity that only for some items “literal” translations are provided, whereas there are more that appear to be polymorphemic by the looks of them but are left unanalyzed in the presentation of the data (though in Section 6.4., the morphological structure of Piaroa body-part terms is discussed in detail).

The first part of the book is concluded by W.D.L. Silva’s “A typology of body-part words in Eastern Tukanoan languages”, which, as opposed to what the title may suggest, actually focuses on possessive constructions rather than words themselves. The basic possessive constructions in most languages involve simple juxtaposition. There is an interesting construction in Koreguaje involving a possessive marker that is prefixed to a classifier (and apparently sometimes also to nouns) to express the possessum, e.g. chʉkʉ́na ne-ñʉ́ ‘our tree‘, ñʉ́ being the classifier for trees. In other languages, juxtaposition is complemented by a strategy involving the marker ya which, as in Koreguaje, can take classifier suffixes. In Desano and Makuna, this situation, involving the same morpheme ya, amalgamates into an alienable-inalienable contrast, whereas in Mʉ̃tẽã, the presence of ya is governed by the animacy of the possessum. The situation in other languages is largely similar, with some minor language-specific differences that can be distilled from the discussion. Generalizations are left mostly to the reader until the summary in Section 7.2.3. This summary is followed by Silva’s claim that juxtaposition is the older strategy, inherited from proto-Tucanoan, which is replaced by strategies involving overt possessive morphemes like -ya. While Silva considers those developments “evidently the results of contact with Arawakan languages” (p. 211), the morphology has internal sources. Specifically, Silva argues that Koreguaje ne is a grammaticalized version of a noun meaning ‘thing’ that is still attested in other languages, and that reconstructs to proto-Tucanonan as *ja, pointing to the observation that “*j is reflected as alveolar sounds in many languages of the family” (p. 208). It is not clear from this statement if these are regular correspondences – which would be important to evaluate the claim – and it is also not entirely clear from the discussion that *ja > ya in the other languages – the spelling confusingly changes back and forth here.

Alexandra Y. Aikhenvald’s chapter “Tariana body parts in North Arawak perspective: What makes a human live?” opens up the second part of the volume. Aikhenvald shows something similar to what Cúneo and Messineo propose for Toba: a basic distinction between internal and external body parts that are associated with different behaviour and that together make up -daki, the material body of a person (the term also describes “peoples’ customs and their way of life”; p. 216). -daki contrasts with -kale, an “invisibly entity [that] is the locus of thinking, feeling, and internal speech” and that also translates as ‘heart’ (p. 216). -kale’s polysemic nature is brought to light clearly in that the material and immaterial senses are associated with different grammatical behaviour. In fact, Aikhenvald is able to trace both readings to two different etyma, *kawale(-ri/ru) and *-ka(a)le, showing that, in this case “the synchronic polysemy … is the result of a historical accident and a conflation of two somewhat similar but originally diverse roots” (p. 221) – an important reminder for both comparativists and descriptivists that apparently related senses might not always be so historically. Aikhenvald provides a very rich and detailed analysis of both -daki and -kale, drawing both on passages of stories and the description of shamanic practice to highlight facets of their nature and their contrast.

Swintha Danielsen and Lena Terhart discuss “Body-part terms in Baure and Paunaka”. Baure and Paunaka are closely related Arawakan languages of Bolivia that show both similarities and differences in how they treat body parts grammatically. For instance, the manner to form inalienably possessed nouns is similar in both languages and operates with clitics/prefixes, but the ways to create derelationalized, unpossessed body-part terms are different both in terms of the morphemes involved and in the token frequency with which such terms occur. Possession can also be expressed through phrasal constructions, and in Baure but not so much in Paunaka, body parts frequently are compounded or incorporated into verbs. Danielsen and Terhart discuss issues of boundaries of terms pertaining to parts of the body that do not have clear natural divisions, prominently the limbs. Engaging with the typological literature, they show that Baure, Paunaka, and other languages of the same region represent different, more complex types of making lexical splits when both arm and leg are given consideration.

Sidi Facundes, Marília Freitas, and Bruna Fernanda Lima-Padovani discuss data from another Arawakan language, Apurinã. In Apurinã, nouns categorized as alienably possessed take a suffix in addition to a possessor proclitic. The latter also appears with inalienably possessed nouns in their possessed form. Inalienably possessed nouns are inherently relational and hence obligatorily possessed, and include, but are not limited to, body-part terms. There is the suffix -txi that can derelationalize them and allow them to be used without an expressed possessor; the suffix -txi, interestingly, doubles as an abstract nominalizer, raising questions as to a possible diachronic relationship. In some ways resembling a situation commonly found in Mesoamerica, in Apurinã inalienably possessed nouns include a subclass called “classificatory nouns”, which are phonologically bound, typically denote plant terms originally, and act as a specifier for shape, texture, and other specifics of the classified referent in a way that can often be described as metaphorical. An example is -tãta ‘bark, shell’, which recurs in the word for ‘eyeglasses’ that combines it with uky ‘eye’.

As Marianne Mithun notes in the introduction to her chapter “Topicality, affectedness, and body-part grammar”, “[i]n descriptions of grammatical constructions involving body-part terms, among the most typical distinctions discussed is that between inalienable and alienable possessions” (p. 287). Mithun argues that in some languages, rather than coding relationships of possession, relevant constructions actually convey something different, namely the topicality and affectedness of referents in the context of the actual discourse. She shows that this second dimension is relevant also to languages without an alienable-inalienable distinction: In languages like Mohawk and Central Alaskan Yup’ik speakers have the choice between using an independent body-part noun or an incorporated one. According to Mithun, this choice allows speakers to code the degree of topicality and affectedness of body parts (and their possessors). In Central Pomo, the dimension of affectedness materializes in a different way. As Mithun observes, pronominal sets and case clitics that can combine with body-part terms are identical to those marking patients and beneficiaries on the clause level. This pattern, too, suggests that the former actually do not encode possession but rather direct and indirect affectedness respectively. In this way, these constructions, in Mithun’s analysis, turn out to furnish speakers with a means to map highly topical participants to core arguments. In the final section, Mithun discusses data from neighbouring Californian languages and argues that similarities in the way argument encoding and possession are intertwined are unlikely to be coincidental, but indicative of language contact. Mithun, furthermore, shows that inalienable possession is possible for some but not all body parts in Mohawk, and that these are “generally those with an exterior surface”, bringing together the question of the grammar of body parts, specifically the alienable-inalienable distinction, with the distinction between interiority versus exteriority once more and adding further facets to it.

Verónica Nercesian and Alejandra Vidal discuss “Body parts and possessive constructions in Mataguayan languages”. In their analysis, the languages encode an alienable-inalienable distinction. “A small number of nominal roots” (p. 310) insert what they analyze diachronically as a prefix of the shape *t(V)- between the root and the head-marking possessor prefix. It is not entirely clear from this phrasing if the terms where the element occurs are all body-part terms, but the interpretation that “this morpheme historically helped distinguish body-part terms from the rest of inalienable nouns” (p. 311) suggests that this is so. Nercesian and Vidal show how this prefix has fused with the stem in some languages, has become part of pronominal prefixes in others, but retains some productivity in Wichí. Their claim that there was originally a suffix of the shape *t(V)- is supported by three independent types of evidence: formal similarity to the free pronouns; the position in the possessed noun that is analogous to a synchronically attested morpheme that codes alienable possession; and the frequency of syllabic reduction and vowel elision in the languages. Regarding the first point, however, the authors then go on to state that “free pronouns, in general, resemble paradigms without t-”. I have tried for a long time to wrap my head around this and came to the conclusion that this must be an argument against previous analyses that would have it that the element in question is part of the pronominal paradigm, and that the point must be that the shape of free pronouns actually lack /t/. In general, comprehension of the argument is complicated for the reader by the way the chapter is organized, in particular the frequent back-and-forth switching between previous analyses and the authors’ own, and repetition of the same points but with different wording that leave one wondering if the same claim is repeated or something new is added to it.

In Chapter 13, Felipe Hasler, Mariana Poblete, Consuelo Sandoval, Felipe Neira, Daniela Aristegui, and Ricardo Pineda discuss “Word-formation strategies and syntactic behavior” associated with Mapudungun body-part terms. This chapter explicitly follows the analytic dimensions of Lehmann’s opening chapter and relies on the combination of a corpus-based approach with elicitation. Compared with the descriptive chapters by Rosés Labrada and Danielsen and Terhart, Hasler et al.’s chapter presents yet another way to structure the discussion. They group Mapudungun body-part terms along their formal properties, starting out with monomorphemic body-part terms (mentioning “coining of simple terms” (p. 340) as the associated strategy, though it remains unclear to what extent, if at all, these terms have been observed to have been “coined” at some point rather than simply inherited). Hasler et al. posit a continuum from transparency to opacity in the language. Some aspects of this may be questioned. For one, the simple terms, which are completely arbitrary, are surprisingly put at the fully transparent pole, showing “absolute semantic transparency, since they are defined non-relationally to other parts of the body” (p. 351); for another, derivatives rank higher than transparent compounds, and the reasons for this are unclear. Finally, like other chapters in the volume, Hasler et al. move on to discussing the grammar of body-part terms in Mapudungun, specifically their syntactic behaviour.

Pilar M. Valenzuela promises to treat “Plant and animal body-part terms in Shiwilu grammar: classification, nominalization, and incorporation”. In Shiwilu (Cahuapanan), body-part (and kinship) nouns can be independent words; a set of plant and animal body-part terms are bound and obligatorily possessed. These items “tend to coincide with the first or the last syllable of semantically related independent or alienable nouns” (p. 375), much like in Panoan languages (the subject of the following chapters). In addition, plant-part terms are reflected in another grammatical system of Shiwilu, namely a set of classifiers that is highly productive in lexical derivation, but that also is utilized in a wide variety of morphosyntactic constructions. Around 75% of the classifiers can be traced to nouns for plant and animal-body parts, and judging from Tables 14.3–14.5, classifiers and inalienable bound nouns are either very similar or, in the majority of cases, identical, indicating, as Valenzuela notes, only a moderate degree of grammaticalization. Unlike Panoan (see the contribution by Zariquiey and Valenzuela), the diachronic relationship between classifiers and alienable nouns might vary, with some classifiers originating from free alienable nouns and others being the source for the free noun. Classifiers are argued to have grammaticalized further into nominalizers, and Valenzuela also dedicates a separate section to the incorporation of body-part terms.

The last three chapters of the volume deal with Panoan languages, specifically the approximately 30 body-part prefixes that are a hallmark of most languages of that family. The first is David W. Fleck’s “Vestiges of body-part prefixation in Marubo”. It deals with a Panoan language in which, as the title suggests, this system has become unproductive. However, it remains visible in as many as 268 nominal stems. Analysis of the non-prefix part of these stems allows the identification of recurrent elements. Most of these are not synchronically attested, but their semantics can be inferred by “subtracting the meaning of the prefix and considering the common meanings of the stems” (p. 415). In spite of this, however, Fleck shows that body-part prefixes are entirely unproductive and lexicalized in Marubo, the main argument being that prefix-root combinations are attested in a scattered manner, and that more complete exploitation of combinatorial possibilities, which other Panoan languages in which body-part prefixes are productive attest, is not in evidence. Furthermore, novel Marubo prefix-root combinations coined by Fleck, while parsable to Marubo speakers who are bilingual in Matses, in which body-part prefixation remains productive, were explicitly rejected.

In “The grammar of body-part expressions in Iskonawa: lexicalization, possession, prefixation, and incorporation”, by Roberto Zariquiey, Jaime Montoya, Juana Ticona, Luz Carhuachín, Yessica Reyes, Roxana Quispe-Collantes, José Paz, and Aarón Torres, the topic of body-part grammar, including prefixes, in Panoan is continued. Like Hasler et al. for Mapudungun, Zariquiey et al. divide body-part terms according to their morphological structure ranging from monomorphemic items, which interestingly cluster in the domain of internal body parts, to morphologically complex forms that include nominalized verb forms and compounds. In addition, there is a large class of etymologically but not synchronically analysable body-part terms that consist of a body-part prefix and sometimes recurring phonological, and probably originally morphological, material. Unlike in Marubo, however, their semantics usually cannot be determined. Zariquiey et al. describe, furthermore, different possessive constructions that, in their analysis, are not related to the alienable-inalienable distinction. Rather, the choice has to do with the referentiality of the possessors.

While diachronic considerations already play a role in the two contributions on Panoan body-part prefixes, the final chapter in the volume, by the editors Roberto Zariquiey and Pilar M. Valenzuela, finally moves entirely into diachronic terrain. At the same time, at least some of the prefixes themselves are seen as having originated from earlier body-part nouns. While synchronically there are no constructions in Panoan languages that might have been the precursors to body-part prefixes on adjectives and verbs, Zariquiey and Valenzuela argue that constructions attested in Tacanan are plausible input to the grammaticalization process that ends with Panoan body-part prefixes. In addition, Iskonawa as described by Zariquiey et al. in the preceding chapter, might support this scenario.

3 Evaluation

In their introduction, the editors state that this volume “will have a significant impact on linguistic typology and cognitive linguistics” (p. 1). There are some general issues that I see with the volume, and ones that I believe are to its own detriment, in particular when it comes to the possible impact that the editors mention. The volume’s internal organization, its coherence, and the visibility of general lessons and insights are less effective than they could be: the title of the volume and the two parts into which it is divided do not fully reflect their contents. Extracting generalizations and red threads that run through individual contributions is hindered by the paucity of cross-references, the absence of a synoptic concluding chapter that would profile recurrent topics in the contributions and what can be learned about them, and at times varying terminology for the same phenomena.

Regarding the first point, it is certainly the case that the contributions in the part titled “Categorialization [sic!], lexicalization, and semantic processes associated with body-part expressions” fall more into the realm of lexical typology, and the chapters in the section “The grammar of body-part expressions” into that of morphosyntactic typology. In this sense, the division has some merit; however, it is not sharp in that the chapters in the first part also discuss the grammar of body-part terms, whereas those in the second part also discuss lexical matters. Furthermore, as the presence of the first part shows, the editors themselves do not consider the volume to be only about the grammar of body-part terms – but this is not reflected in the title.

One quibble related to the second point is the poor integration of Lehmann’s opening chapter with the rest of the volume. It is cross-referenced in some chapters, but these cross-references usually pertain to very well-circumscribed points. As far as I could see, only Hasler et al.’s chapter on Mapudungun takes up the “foundations” laid out by Lehmann at large. This leaves Lehmann’s chapter, which by its position and content gives the impression of providing a basic analytical grid for the contributions in the volume, very much standing alone. Relatedly, many contributions insightfully discuss topics and analytic problems in the lexicon and grammar associated with body parts across American languages. Many of these topics are relevant in more than one contribution. For instance, the central topic of Mithun’s chapter is that affectedness and topicality are dimensions that are relevant to the syntax of body-part terms. This very point recurs in both the chapters of Hasler et al. on Mapudungun and Valenzuela on Shiwilu (and indeed, noun incorporation, which Mithun shows to be the relevant morphological process to regulate the status of a body-part term along these dimensions in languages like Mohawk, is also a relevant process in both Mapudungun and Shiwilu). Valenzuela, like Mithun, even explicitly brings into play the “possessor raising” construction – yet unfortunately, there are no cross-references that make this common topic across contributions (more) visible to the reader.

There are other more general questions that the rich material in this volume raises. What is the panorama of variation regarding the dimension of possession in the grammar of body parts? Possession is addressed in most chapters, and there is variation here. But the dimensions that the chapters carve out are not related to one another, nor to Lehmann’s chapter, and there are no cross-references that would point out similarities and differences, leaving the task of comparison to the reader. How do body-part terms interact with classifiers in languages that have them? The same remarks apply. What are the ways in which the grammar of internal and external body parts can vary? Again, this is a topic that recurs in several contributions, but this observation is very much an accidental find for which one has to sift the entire volume. Finally, given that this volume hones in on languages of the Americas, what is specifically “American” among the phenomena contributors describe? Such questions might have been addressed by a possible closing chapter that brings together the evidence from the original contributions. Comparability is also reduced for the readers because the same topics are discussed in different orders in many individual chapters. For instance, Rosés Labrada presents first a catalogue of body-part terms, divided into parts that are consistent with Piaroa categories, and then moves on to the discussion of their grammar, whereas Danielsen and Lenhart choose the opposite direction, and present terms not by body regions, but etymological status. This would be less remarkable if the information provided were better cross-referenced and tied back to a basic analytic grid such as that provided by Lehmann.

A third issue is that definitions of core notions like alienability and inalienability, like any other on which much has been written, differ in the literature. Indeed, different chapters provide definitions of what they take these notions to mean, but there does not seem to be one that is valid for the volume as a whole, as established by editors, for instance. I have not checked whether variation in how terms are defined affects comparability across chapters, but I do see at least the danger that it might, even if only in minor ways. Terminology is also a problem elsewhere. For instance, Lehmann (p. 26) speaks of “derelationalization” to refer to the processes by which an obligatorily possessed (body-part) noun can be made to appear without a possessor. But in their contribution on Apuriña, Facundes et al. speak of “unpossession marking” even though obviously the very same process is implied.

Prima facie, thus, the volume would be useful to scholars interested in a specific language, who might read the chapter on that language and bypass the others; someone interested in metaphorical extension of body part terms might read Bowers’ contribution and perhaps others, but missing the information contained in those chapters that are not profiled in the introduction as containing discussion of relevant processes; someone interested in more anthropologically oriented questions might read Aikhenvald’s chapter, etc.

For sure there are interesting and new things to be learned in the chapters assembled in this volume. Many of the descriptive chapters in particular are masterful. I am glad that I have read the volume, and I hope to have shown here how the wealth of information the volume provides can also inform broader questions as to the organization of body-part terms as a semantic field and especially the grammar associated with them – for those who devote the time to studying the volume in detail, as it surely deserves.