Zariquiey, Roberto, Masayoshi Shibatani and David W. Fleck: Nominalization in languages of the Americas

1 Overview

A small clarification to start with: Despite its title, this book only deals with so-called Native languages of the Americas – i.e., not colonial languages or creoles – and more specifically, with languages of South America. The volume is the outcome of a workshop with the same title, which took place in Lima in 2014.

It has often been observed that nominalization is the main strategy in the formation of complex sentences in Native American languages. In the introduction to the volume (pp. 1–11), the editors state that despite this fact, nominalization is often discussed from the perspective of subordination and only rarely appreciated in its own right. They intend to remedy this situation, following Vendler’s (1967: 125) insight that “the grammar of nominalizations is a centrally important part of linguistic theory” (p. 2). Consequently, the subsequent studies on individual languages present nominalization patterns and their various functions, in contrast to the more common approach that starts with an analysis of subordination patterns and finds that subordination is often realized through nominalization.

After an introduction by the editors, the volume is divided into three parts, containing 17 chapters altogether. Part I, “Nominalization theory” and Part II, “Areal studies on nominalization in South America” each consist of only two chapters. The remaining 13 chapters are assembled in Part III, “Case studies on nominalization in individual languages”. I discuss the two theory chapters of Part I in Section 2 and provide an overview of the chapters of Parts II and III in Section 3, closing with a general evaluation of the volume in Section 4.

2 Part I: Theory and diachrony

Masayoshi Shibatani’s chapter “What is nominalization? Towards the theoretical foundations of nominalization” presents a functional approach to the concept of nominalization, richly illustrated with examples from languages from all over the world. Shibatani (henceforth MS) recapitulates the distinction between lexical nominalization, which creates new lexical items, and grammatical (also known as clausal or syntactic) nominalization, which turns clausal structures into nominal constituents. While there are many overlaps between the two, it is this latter type that MS is most interested in. According to MS, structures traditionally treated as relative, complement or adverbial clause, as well as classifier and possessive constructions can all be characterized in terms of grammatical nominalization. Under this view, what is traditionally termed relative clause is an argument nominalization, which can be used in referential (when headless) as well as in modifying (when headed) function; complement clauses are event nominalizations in referential use; and adverbial clauses are event nominalizations in modification use. In contrast to more traditional analyses, although they generally acknowledge that dependent structures can be identical to or have developed out of nominalization, MS argues that these structures are nominalizations in all languages, and that the label “clause” for these structures should be dispensed with altogether.

This radical position is made possible by a definition of nominalization that does not take into account structural properties. MS defines nominalization as “a metonymy-based grammatical derivation process yielding constructions associated with a denotation comprised of entity (thing-like) concepts” (p. 21). Structures known as relative, complement, or adverbial clauses are nominalizations because they “denote”, namely “events, facts, propositions, resultant products, and event participants” (p. 21). A clause, in contrast, does not “denote”, but “predicate”, which, according to MS, is not what relative, complement, or adverbial clauses do.

In MS’s theory, nominalization does not have to be morphologically marked. That a nominalization has taken place can be indicated by “NP-use markers”, as in the/a cook, where the determiner identifies the word cook as a nominalization. Also the syntactic position can signal nominalization. So, the structure John is honest is a clause when uttered independently, but within the sentence I know (that) John is honest it occupies the position of a noun and hence denotes a “fact” (pp. 72, 93). Furthermore, in contrast to many other accounts, MS allows for nominalization to be based on nouns, as in the case of English villager, which is derived by attaching the nominalizing suffix -er to a nominal base.

The idea – already formulated, in one way or another, by Vendler (1967), Hopper and Thompson (1984), Langacker (1987, 1991 – that the information expressed in dependent structures tends to be conceptualized as “thing-like”, is quite plausible. However, the extent to which it is applied here is not unproblematic for descriptive or typological analysis.

For instance, there is some circularity in the argumentation that nominalizations are identified on the basis of their structural properties (e.g. syntactic position), but that they have these structural properties because of their meaning. This is at least how I understand passages like the following: “As products, nominalizations are like nouns … by virtue of their association with an entity-concept denotation, a property that provides a basis for the referential function of a noun phrase headed by such nominalizations” (p. 19, emphasis in original). Or later, with regard to argument nominalizations: “[N]ominalization structures function as syntactic arguments as a subject or object precisely because they evoke and stand for thing-like entities just like ordinary nouns do” (p. 72).

Another problem is MS’s insistence that nominalizations can be derived from nominal bases. This makes me wonder where the line is drawn between nominalizing and nominal derivational morphology. For instance, the Dutch diminutive can have a nominalizing function, as in uit-je (out-dim) ‘(the/a) stroll’, but arguably it is still a nominal suffix. I also doubt that it is always clear where to draw the line between primary nouns and zero-marked noun-based nominalizations. MS here recurs to his definition of nominalization as a metonymy-based process: So, the noun violin in the sentence The first violin is sick today (p. 72) is a nominalization, since the NP refers to the person playing the instrument rather than to the instrument itself. However, the definition is semantically based and therefore difficult to grasp.

Also regarding deverbal nominalization, MS’s theory is only of limited use for identifying nominalizations. The example of cook works well, since in an NP it denotes an entity defined in terms of an activity, in analogy with agent nominalizations derived with -er. But how would the theory, which does not take into account diachronic developments, deal with words like iron (the instrument used for ironing) or plant (in the biological sense), which can also be used for both predication and reference? Would they be considered unmarked nominalizations when used as nouns? As verbalizations when used as verbs? As primary nouns? or as pre- or acategorical – the lexical category being attributed according to the syntactic environment? These possibilities are not addressed in the paper.

Still more challenging to the theory are syntactically flexible languages, i.e. languages in which both nouns and verbs can be used for both reference and predication, such as Tagalog, Salishan languages, Classical Nahuatl (Launey 2004), or the language of my own study, the South American isolate Movima (Haude 2019). I assume that in MS’s theory, in such a language the occurrence of any lexical item inside a referential phrase would be considered a nominalization, including that of a noun. However, the notion of nominalization becomes void in such cases, where the functions of predication and reference are systematically dissociated from lexical categorization (a point also made by Cruz and Praça in this volume, see below). It is also impossible to say for sure that (verbal or nominal) predicates in such languages do not “denote” “entity (thing-like) concepts”, a capacity that MS only allows for nominal expressions.

In sum, MS’s proposal contains inspiring ideas, illustrated with data from a wide range of languages. Still, there are open questions, and it would be nice to see this chapter expanded to a full monograph including a more thorough discussion of the aspects neglected or not fully developed here.

The chapter contains many typographical errors, including misspelled names of languages (“Salve” for Slave, pp. 30, 31, 101) and authors (“van Dijk” for van Gijn, p. 75), as well as obviously unintended terminology (“content clause” for “complement clause”, p. 74; “event nominalization” for “lexical nominalization”, p. 95, end of Section 6.3). The glosses are often inconsistent (e.g. the alternations between ‘msc’ and ‘masc’ or between ‘s’, ‘sub’, and ‘subj’). The list of abbreviations is dramatically incomplete, explicitly containing only “less commonly used abbreviations” (p. 160). Rather than being based on some standard like the Leipzig Glossing Rules, the choice is quite arbitrary: For example, f ‘feminine’ is listed, but n (presumably for ‘neuter’) is not. Even the bracketed sections of the examples, which contain the nominalized segments under discussion, contain undefined glosses. For instance, ‘nom’ (particularly confusing in the Mayrinax Atayal example on p. 48, whose glosses do not reflect the analysis in the text) presumably stands for ‘nominative’, but is also quoted as an abbreviation for ‘nominalization’ (pp. 31, 116). The reader is left with a lot of guesswork, which could have been avoided if the (co-)editors – of the volume and of the book series – had been more attentive.

In the following chapter, “Nominalization in cross-linguistic diachronic perspective”, Sonia Cristofaro (henceforth SC) concentrates on the origins of nominalizations. This topic, she states, has been poorly explored – in contrast to the further developments that nominalizations undergo and that have been the subject of many diachronically oriented studies, including some in this volume. SC resumes accounts in the literature showing that nominalizations usually originate in lexical elements with a generic meaning (‘thing/matter’, ‘person’) in a referring function, or from referring elements like demonstratives. In the source construction, she argues, these elements are combined with modifying or predicating expressions. As the meaning of the referring or entity-denoting expression gets bleached over time, it develops into a nominalizer.

SC contrasts this diachronic perspective with typological-functional accounts (Croft 1991, 2001; Hopper and Thompson 1984) that consider nominalization as marking a non-default employment of a process/state-denoting expression. SC argues that the disappearance of verbal categories is due to the fact that the expression does not “describe a specific occurrence of some state of affairs” (p. 178), and that the addition of nominal ones (such as case, number, or gender marking) that characterizes nominalizations is actually a residue of the originally referring expressions. Rather than simply marking a non-default employment, SC argues, the diachronic pathways in which nominalizers emerge are as diverse as the resulting structures. For instance, from a functionalist perspective it is difficult to explain why in one language certain non-default uses of an item are overtly marked and others not; usually, the explanation lies in diachrony.

It is indeed puzzling that, as SC points out, it is so easy to detect developments that nominalizations undergo when evolving into finite structures, but that it is so difficult to identify a structure that is developing into a nominalization; SC’s addressing this point is an important addition to the topic. It is not entirely clear, however, why SC considers the diachronic approach to stand in such stark opposition to discourse-functional explanations. It might be conceivable that the original constructions leading to nominalization – e.g., the combination of an event-denoting word with a referential element – is used to signal the non-prototypical employment of that word. SC mentions this possibility in a footnote (p. 177), but rejects it as “irrelevant”, since, she argues, the original construction does not contain a nominalization. This position seems to contradict Shibatani’s, who would probably consider the referential elements “NP-use markers”, which signal that a nominalization is involved. In any case, it is difficult to imagine what the original construction might have looked like – the paraphrases of source constructions in this chapter all involve derived forms in English, e.g., “‘matter/thing (of) Verbing’” (p. 173).

3 Part II and III: Nominalization in Native (South) American languages

The papers in Part II and III of the volume provide ample evidence that in the languages of the Americas, relativization, complementation and adverbialization are indeed usually realized through nominalization. Many chapters also take into account the diachronic dimension, showing how different types of nominalization develop from one another and into new grammatical structures. It is impossible here to do justice to the careful analyses presented in each of these chapters. My comments are merely spotlights on some central points they make.

The two chapters in Part II take an areal perspective and are largely based on data from the literature. The contribution by Rik van Gijn (henceforth RG) “Case markers as subordinators in South American Indigenous languages” is an excellent overview of the distribution of case marking and their extension to subordinate clauses in languages of South America. With nominalization as the principal means for forming subordinate clauses in South American languages, case markers – in languages that have them – serve to mark different types of subordination, and they usually follow the typical functionally expected patterns. For instance, accusative case is extended to the marking of complement clauses (especially in Quechuan), spatial cases (locative, ablative, inessive) are extended to mark temporal adverbial clauses. The study is based on a sample of 60 languages representing 26 families and 10 isolate languages. Accordingly, the results are highly diverse, but carefully structured and analyzed. With 18(!) maps, useful tables and figures, and summaries of every topic treated, the paper is a valuable source of information for anybody interested in subordination and case marking in the region. Slightly contrasting with Cristofaro’s position, RG suggests that diachrony is only one of the factors that contribute to the shape and function of a particular construction, and that cognitive principles may play a role as well.

Lucía A. Golluscio, Felipe Hasler and Willem de Reuse discuss “Nominalized constructions with argument functions in languages of the Chaco”, i.e. the languages spoken in the geographical area between south-eastern Bolivia, northern Argentina, and part of Paraguay. They concentrate on 10 languages equally distributed over five families (Mataguayan, Guaycuruan, Zamucoan, Tupí-Guaraní, and the near-extinct Lule-Vilela) and compare them with standard characterizations of the neighboring Amazonian and Andean languages.

A noteworthy result of this study is that in contrast to Andean and many Amazonian languages, in Chaco languages subordination is formally unrelated to nominalization. Rather, nominalized expressions with argument functions are expressed by finite forms with clause-initial subordinators, sometimes also by paratactic constructions. The authors reconcile this finding with Shibatani’s approach by regarding nominalization as an exclusively functional phenomenon: They state that the subordinate constructions indicate dependency and that it is “the fulfillment of this function, and not the presence of a special subordinator, that indicates nominalization” (p. 253).

The 13 chapters in Part III deal with either one single or a few related languages. They are for the most part based on first-hand data from the authors’ own fieldwork, including a large amount of corpus data of spontaneous oral discourse. As is common in this kind of collection, the authors either give an overview of all aspects of the phenomenon in their languages of study, or they concentrate on one or more particularly noteworthy aspect.

Most papers address the similarities and differences between so-called lexical and grammatical (or clausal, or syntactic) nominalizations. At first sight, the distinction seems clear: lexical nominalization derives new lexical entries, whereas grammatical nominalization derives dependent structures from clauses and can also be used in an ad-hoc way to form what are known as relative, complement and adverbial clauses. However, in many languages the same derivational process can be used to form both lexical entries and ad-hoc formations. Similarly, also the apparently clear-cut distinction between relativization, complementation, and adverbialization can show remarkable overlaps.

The similarity between lexical and grammatical nominalization is demonstrated in the first chapter of this section, “Nominalization in Central Alaskan Yup’ik” by Yuki-Shige Tamura. For instance, in Yup’ik the term ‘our policeman’ is literally expressed as ‘the one who took us’ (p. 275) and has the same structure as any other argument nominalization. Furthermore, these nominalizations are used as modifiers – equivalent to headed relative clauses – by simple juxtaposition.

This potential of participant nominalizations to be used for modification, i.e. for restrictive relativization, through simple apposition is treated in more depth in the following chapter, “The ‘relative’ illusion and the origin of non-subject nominalizers in Cahita (Uto-Aztecan)” by Albert Álvaro González. In the Cahita languages (Yaqui, Mayo, and the extinct Tehueco) of northwest Mexico, object nominalizations originate from three different sources: possessed classifiers, subject nominalizers, and spatial postpositions. All these elements used to be noun phrase markers, which were first used referentially and later as appositions. That is, the modifying function is a secondary development from the referential function, a proposal already made by Givón (2012) and refined by the author of this chapter.

In the chapter “On habitual periphrasis in Cuzco Quechua”, Rammie Cahlon argues that the so-called Past Habitual Periphrasis in Quechua is constructed from an agentive nominalization combined with the copula (in the case of the third person, zero). Based on published sources, the author shows how this construction acquired its present-day meaning.

The contribution “Life of =ti: Use and grammaticalization of a clausal nominalizer in Yurakaré” by Sonja Gipper and Foong Ha Yap is a corpus-based analysis of a nominalizing clitic that attaches to fully inflected predicates to form relative, complement, and adverbial clauses – with sometimes ambiguous interpretations. Despite the lack of historical data on this linguistic isolate, the authors convincingly argue that the nominalizer originated in the demonstrative ati, and that its initial argument-nominalizer function was extended to the modifying, complementation and adverbial functions, thereby acquiring a more and more abstract meaning. In addition, a nominalization with =ti may function as an independent sentence. As the authors explain, these “syntactically incomplete” constructions are anchored in discourse by various non-syntactic strategies such as contextual information through prior interactions between the interlocutors. In some of these stand-alone instances, the nominalizer =ti has been reinterpreted as a clause-final enclitic marking the speaker’s intersubjective commitment.

In “The rise of the nominalizations: The case of the grammaticalization of clause types in Ecuadorian Siona”, Martine Bruil discusses the development from classifiers into nominalizers, from nominalizers into subject agreement markers and from these into markers of reportative and interrogative clauses in a Tucanoan language. In this way, the nominalized complement clause ‘They say that he studies’ has acquired the reading ‘He studies (they say)’ (p. 407), in which the masculine nominalizer has been reanalyzed as a masculine subject marker.

The chapter “Form and functions of nominalization in Wampis” by Jaime Peña is a detailed description of suffixes that derive lexical nominalizations, but that are also used for creating relative, complement and adverbial clauses. Also in Wampis, a Jivaroan language from Peru, nominalizations can expand to main-clause predication, first by being combined with a copula and then by receiving finite verb morphology. Often, these stand-alone nominalizations are used to mark reported speech or surprise. Sometimes, nominalized forms are even treated like verb stems, with verbal inflectional morphology added to them.

In the chapter “Nominalization in Harakmbut”, An Van linden describes two prefixes marking participant nominalization (called “argument nominalization” in most other chapters) and event nominalization, respectively, in a linguistic isolate from Peru. Here as well, there is no clear separation between lexical and grammatical nominalization. Participant nominalizations can head NPs, but can also modify other nouns. The author notes that Harakmbut also has a dedicated relative-clause marker based on finite verb forms, but apparently she does not consider this a nominalization (p. 464, fn. 5). One nominalizing prefix can also derive independent nouns from bound forms, is found in compounds, and can derive independent nouns from inalienably possessed ones.

Luis Miguel Rojas-Berscia’s contribution “Nominalization in Shawi/Chayahuita” differs from the other chapters in this section in that it concentrates on the assessment of theoretical frameworks for the analysis of nominalization patterns. The author chooses Malchukov’s (2006) Generalized Scale Model for describing selected deverbal nominalization patterns in Shawi, a Kawapanan language from Peru. In some of these nominalizations, all or most verbal features are lost (Malchukov’s “strong nominalization”), while in others, verbal features are retained (“weak nominalization”).

The following three chapters deal with Panoan languages, spoken in the Peru-Brazilian border area and in which nominalizations are said to be particularly frequent. The chapter “Clausal nominalization in Kakataibo (Panoan)” by Daniel Valle and Roberto Zariquiey discusses grammatical nominalizations in two Kakataibo dialects. As is typical of grammatical nominalizations, they combine external nominal features with internal clausal properties. The same nominalizing morpheme can derive an argument or an event nominalization, the only difference being that in the latter, there is no gapping of arguments. Hence, when an argument is gapped, the clause is interpreted as an argument nominalization: ‘I saw (the things) that our ancestors did a long time ago’ (p. 524, ex. 18); but when the argument is overtly expressed, the clause is interpreted as an event nominalization: ‘I saw that/how our ancestors did the things a long time ago’ (p. 525, ex. 19). Kakataibo nominalizations can also be used for modification, and they can also occur independently. The authors of this chapter also demonstrate the historical relationship between nominalization and switch-reference and point to the coexistence of both interpretations (p. 531).

The switch-reference marking function of nominalizations is also the topic of the chapter “Nominalization and switch-reference in Iskonawa (Panoan, Peru)” by Roberto Zariquiey. The near-extinct Iskonawa has an impoverished switch-reference system and resorts a lot to nominalizations in order to mark same- or different-subject relations. This can lead to ambiguous interpretations. For instance, the same sentence containing an argument nominalization can be interpreted as ‘Germán, whom I hit, cried’ or as ‘After I hit Germán, he/someone else cried’ (p. 550). Grammatical nominalizations in Iskonawa are argued to be often formally indistinct from finite clauses. However, it seems that nominalizations tend to involve a copula, which can be interpreted as signaling the semantic shift from an event- to an entity(like)-denoting interpretation.

In the chapter “Lexicalized nominalized clauses in Matses (Panoan)”, David Fleck presents the frequent identity of lexical and grammatical nominalizations as a lexicographic problem: How to deal with expressions that one might analyze as relative clauses, but that are used as lexical items in a language? For instance, the word for ‘broom’ is literally ‘sweeper’ in Matses, a form that is also used in ad-hoc formations like ‘what is/will be used for sweeping (e.g. a branch)’ (p. 561). These argument nominalizations can be rather complex and may, for instance, include adverbial expressions.

The following chapter, by Adriana Estevam, deals with “Nominalization and its pervasiveness in Xavante”, a Jê language of Central Brazil. This study describes lexical and grammatical nominalizations and includes a diachronic hypothesis on the rise of independent clauses out of nominalizations. Unlike the other languages studied in this volume, Xavante is unusual in that lexical and grammatical nominalization are carried out by distinct morphology: Lexical nominalizations are derived by suffixes, whereas grammatical nominalizations are marked by a prefix. The author suggests (p. 601, fn. 7) that the prefixes used for lexical nominalization may have arisen from nouns, while the nominalizing prefix may have evolved from a demonstrative. Nominalized clauses can also be used as independent clauses expressing sentence focus. Even finite clauses in Xavante exhibit traits of nominalizations.

The last chapter, “Innovation in nominalization in Tupí-Guaraní languages: A comparative analysis of Tupinambá, Apyãwa and Nheengatú” by Aline da Cruz and Walkíria Neiva Praça is the only contribution that includes “omnipredicative” languages –Tupinambá and Apyãwa – i.e. languages in which both nouns and verbs are basically predicates and must be overtly marked in order to function as arguments. Rather than following Shibatani’s theory (see above), the authors argue that there is no nominalization involved here, asking “[W]hat would be the use of nominalizations in these languages?” (pp. 630–631). The innovative Nheengatú, by contrast, has lost the omnipredicative pattern. Here, words denoting verbal and nominal concepts receive a distinct syntactic treatment, and nominalization signals the referential use of a word denoting an event, process, or state (in line with Croft 2001, discussed by Cristofaro, see above).

4 Evaluation

This volume is a highly welcome addition to the studies of nominalization in languages from different regions of the world (see the comparable volume edited by Yap et al. (2011) on Asian languages). The descriptive chapters are a treasure for anyone interested in the strategies Native American languages employ to form dependent structures, both from a synchronic and from a diachronic perspective. Despite the concentration on South American languages, many of the patterns described here are also found in languages of Central and North America, so this book can also serve as a basis for the description of subordination and nominalizations in Native languages in other parts of the continent.

The volume shows that the formation of relative, complement, and adverbial clauses in the way known from Indo-European languages, in which (near-)finite structures are introduced by some conjunction and which coexists with participles and lexical nominalizations, is uncommon in these languages. Rather, dependent structures are usually created through nominalization, supporting the idea that backgrounded states and events are conceptualized as thing-like entities. The advantage of the nominalization approach is that it allows a more straightforward description of structures that seem to have different functions – relativization, complementation or adverbialization – but share many traits that would not be adequately captured if described from the perspective of these functions, as is traditionally done.

Many chapters comment on the diachronic developments of nominalizations, with respect to both the rise of nominalizers and the grammatical markers they can develop into. They confirm earlier findings that nominalization strategies develop out of demonstratives or elements with lexical content, such as classifiers. In several contributions it is argued that modifying argument nominalizations (i.e. headed relative clauses) develop only after referring ones (i.e. free relatives). They also show how nominalized forms further develop into markers of aspect, tense, person, switch reference, and ultimately, into main-clause predicates, the latter often starting out as pragmatically marked structures.

Overall, the volume is coherent in that patterns that are traditionally analyzed from the perspective of subordination are regarded from the perspective of nominalization. Each chapter says something on the distinction between lexical and grammatical nominalization and provides information on the differences and overlaps between different types and functions of nominalizations. For any reader interested in particular aspects of the topic of nominalization and subordination – both formal and functional – it will be easy to retrieve the relevant information for any of the languages treated here. Moreover, and this can be attributed to the care of the editors, there are numerous cross-references between the different chapters in the volume.

Shibatani’s chapter stands out a bit. It is presented as the theoretical backbone of the volume, but the descriptive chapters show that its approach is difficult to put into practice. On the formal side, with 152 pages the length of this chapter is disproportionate. This could have been avoided if the chapter had concentrated on Native (South) American languages, in line with the overall topic. (The regional focus of the volume is underlined by the index, which only lists Native American languages.) I must add that the depreciatory tone Shibatani employs with regard to authors whose analyses or definitions deviate from his own is inappropriate in a volume that seeks to advance our knowledge of lesser-described languages.

The editorial quality of the individual chapters is good, with some exceptions (see my comments on Shibatani’s chapter in Section 2; Cahlon’s chapter contains some typos and typesetting errors; Bruil’s chapter is inconsistent in some places, as if examples had been removed during the revision process). It is not clear whether there was a systematic reviewing process: Anonymous reviewers are acknowledged in only four chapters; half of the chapters do not contain any acknowledgments at all.

These drawbacks do not call into question the overall value of the volume, however. While the first two chapters are of interest especially to typologists and theoretically oriented linguists, the descriptive chapters of Part II and III are an excellent source of knowledge on Native languages of the Americas. They are rich in original data, which are well presented and analyzed. The terminology is generally treated carefully by the authors, differentiating well between different types of nominalizations and pointing out different degrees of finiteness.

Thus, with its large amount of original data, the high quality of description of under-documented languages, and its focus on a phenomenon that is often buried under several layers of analysis, this volume is a highly valuable addition both to the literature on Native American languages and on subordination phenomena in general. Nominalization has often been noted to be a pervasive trait of these languages, and this volume provides excellent empirical evidence of this impression.