In this chapter I examine the phonetic origins of voiceless sonorants cross-linguistically within the general framework of Evolutionary Phonology (Blevins 2004; 2006; 2008b; 2015). In terms of a general hierarchy of contrast, we observe that: voiceless obstruents are common; voiceless sonorant consonants are uncommon; voiceless vowels are extremely rare. One phonetic source of voiceless sonorants is coarticulation in RH and HR and clusters, where R is a sonorant and H is a segment produced with a spread glottal gesture. Voiceless sonorants may also arise when laryngeal spreading gestures are associated with prosodic domains. In this second case, voiceless sonorants can arise as allophones of their voiced counterparts. While a fair number of languages show voiceless sonorant glides, liquids and nasals phonologized as a consequence of RH/HR coarticulation, voiceless vowels resist phonologization despite their high frequency as phonetic variants of modal vowels. In some cases, voiceless vowels are lost before phonologization can occur. In other cases, resistance to phonologization may be due to effects of analogy, /h/, word phonotactics, or lexical competition.
The Obligatory Coding Principle accounts for the inventories of possible coding frames in languages that, according to the current terminology, can be characterized as consistently accusative or consistently ergative in their system of argument coding. In coding frame inventories fully consistent with the Obligatory Coding principle, every coding frame includes a given type of coding, either A (in obligatory A coding languages) or P (in obligatory P coding languages). However, languages with coding frame inventories violating this principle are not exceptional. This chapter examines the types of evolutions that may result either in global shifts affecting the Obligatory Coding Principle, in systematic violations of the Obligatory Coding Principle, or in the gradual spreading of non-canonical coding frames. The idea underlying this study is that, before discussing the theoretical status of this kind of generalization, it is crucial to clarify its involvement in diachronic processes.
A central issue in typology is the role of implicational hierarchies in shaping individual languages. One view is that the hierarchies guide language change, or at least constrain it: “Since a hierarchy constrains what is a possible language, it is also a constraint on language change, because languages move from one possible state to another” (Corbett 2011). Other approaches take a different perspective: “Hierarchies simply capture the outputs of independent diachronic processes” (Cristofaro & Zúñiga this volume). Here the relationship between typology and diachrony is examined with respect to the most frequently-cited hierarchies, the cluster of Referential/Topicality/Animacy/Empathy hierarchies. While such hierarchies might appear to drive diachronic development in some single-step changes, multi-step developments are a different matter.
Sahaptin and Nez Perce, the two languages of the Sahaptian family, have both been cited as case studies in the typological literature on hierarchical patterns in main clause grammar. Nez Perce has ergative case marking on only third person singular transitive subjects, plus a minor pattern of indexation of SAP participants via (rarely occurring) second position enclitics. Sahaptin has one of the more complex hierarchical systems ever described, with SAP indexation via enclitics, third person indexation on verbs, differential object marking, an inverse verbal direction prefix, and two distinct ergative suffixes, each restricted to a subset of third person singular transitive subjects (one when objects are SAP, the other when objects are third person). This paper begins by reviewing, evaluating, and occasionally expanding on existing knowledge: we summarize the hierarchical patterns in Sahaptian and characterize each distinct construction. Then we compare relevant Sahaptin morphemes with cognates in Nez Perce, and review their reconstruction to Proto-Sahaptian. The primary contribution of this paper is organizing the morphemes (and their accompanying hierarchical patterns) in both languages into cognate constructions, then reconstructing each to its Proto-Sahaptian origins. We conclude by reviewing and evaluating proposals for Pre-Proto-Sahaptian developments claimed to explain the origins of hierarchical patterns that reconstruct to Proto-Sahaptian. The mechanisms we identify as having created the Sahaptian hierarchical effects are diverse, some motivated and others not, some arising from internal sources, others arguably from contact.
A set of Old Irish clitic person markers may co-index subject or object markers on the verb. Only one clitic of this set may appear at a time, however. The clitic, if present, indexes the verbal argument highest on the scale 1st person > 2nd person > 3rd person animate > inanimate. The restriction of one clitic per verb, a prerequisite for the hierarchy, is explained here as a result of the deictic origin of the clitic set. The deictic origin further favors the marking of local persons over non-local persons. For the upper and lower end of the hierarchy, however, the ranking appears to be determined rather by considerations of function. Since the clitics can mark topics, they favor 1st persons over 2nd and animates over inanimates.
This paper reconstructs the history of a set of innovated 1st and 2nd person verbal prefixes in Reyesano which manifest the phenomenon of ‘hierarchical agreement’ in transitive clauses, according to a 2>1>3 hierarchy. I argue that these prefixes come from independent ergative-absolutive pronouns which first became case-neutral enclitics in 2nd position in main clauses and then verb prefixes. And I show that the hierarchical effects that the prefixes manifest in synchrony have nothing to do with the working of a hierarchy during the grammaticalization process. In doing so, the paper contributes to the growing body of diachronic evidence against the idea that the person hierarchy is a universal of human language reflecting a more general principal of human cognition.
In this chapter, we extract general principles of language change from the study of the evolution of the conjunct order in various Algonquian languages, and propose four generalizations concerning the directionality of the spread of analogy in these systems. These generalizations are expected to bring insights on the analysis of data from other language families with direct/inverse marking but insufficient philological record, such as for instance Sino-Tibetan.
Tupi-Guarani languages are supposedly perfect examples of hierarchical indexing systems, where the relative ranking of A and P on the 1 > 2 > 3 person hierarchy determines the selection of the person markers. This chapter questions the relevance of the person hierarchy as a synchronic and diachronic explanation for such systems, with data from 28 languages. First, only SAP > 3 can really be posited in the actual languages, and second, it explains only part of the facts that it is supposed to account for in Proto-Tupi-Guarani. The chapter ends by suggesting that these systems do not result from the person hierarchy as a functional motivation. Instead, they may result from grammaticalization of pronominal paradigms lacking third-person forms.
There are three distinct transitive constructions in four Coast Salish languages, Squamish, Halkomelem, Klallam and Lushootseed. In the V-tr construction, both A and P are unmarked for case; in the V-mid construction (often considered antipassive), A is unmarked and P marked; in the V-tr-mid construction (often considered passive), P is unmarked and A marked. Individually, none of these constructions is a hierarchical system, but in combination, asymmetries in their distribution are well on the way to creating a person-based hierarchical system. This paper discusses the diachronic development of each of these constructions, then describes their differential distribution into the four functional domains: local (SAP A → SAP P), direct (SAP → 3P), nonlocal (3A → 3P), and inverse (3A → SAP P). While the distribution is not identical in each of the languages, the trend is clear: the etymologically passive V-tr-mid construction cannot occur in the direct domain and has become the pragmatically unmarked construction in the inverse domain, whereas the etymologically antipassive V-mid construction cannot occur in the inverse domain. While it only occurs in the direct and nonlocal domains, even there it is rare, giving the appearance that its function is still that of an antipassive. In combination, the result is that whenever the two core arguments of a clause are an SAP and a third person, regardless of grammatical role the SAP participant is always an unmarked core argument, whereas the third person is most often marked, leading to a situation where the oblique case in these languages is beginning to resemble the obviative case-marker of inverse languages.
The study of hierarchical argument indexation systems shows that while the ranking of both 1st and 2nd person over other arguments is robust and reliable, it is impossible to find any compelling crosslinguistic evidence for one or the other ranking of the two Speech Act Participants, and rare to find a consistent ranking even within a single language. This paper assembles and reviews historical changes in the indexation of the “local” categories (1→2 and 2→1) in a number of Tibeto-Burman languages. We see that the fundamental deictic ranking SAP > 3 is conservative, and inverse marking to emphasize that ranking has been reinvented several times in the family. Changes in the marking of local categories are more diverse, but two phenomena recur independently in different languages and branches: a tendency for the 1→2 form to be uniquely marked, sometimes with forms which are not synchronically relatable to anything else in the paradigm, and a contrasting tendency for the 2→1 form to merge with the marking of 3→1. I propose that these tendencies reflect what I call sociopragmatic effects, i.e. the socially delicate nature of any and all natural utterances involving both the speaker and the addressee.
The Animacy Hierarchy (AH) is an important construct employed for the description and explanation of variation and splits in case marking and agreement in various grammatical domains. The AH is a scale that combines person, definiteness and semantic animacy and is used to state clear preferences of certain morphosyntactic coding types over others. One assumption of the AH is that proper names (PNs) occupy an intermediate place between personal pronouns and common nouns. Despite the large body of research since its first extensive formulation in Silverstein (1976), it is astonishing that there has been almost no empirical evidence published for this claim. Since the AH has been formulated mostly on the basis of case marking and agreement phenomena in languages with split ergativity or hierarchical alignment, we compiled a sample of more than 30 such languages in order to find data on the morphosyntactic coding of PNs. While there are only a very few instances that confirm the claim, there are more instances that contradict it. We concluded that PNs should be removed from the AH, since their assumed position has no predictive value for typological generalizations.
This paper discusses the history of generic person marking systems in several Gyalrong languages. While closely related, Japhug and Tshobdun differ considerably: the inverse prefix marks generic A in Japhug, while it appears in the generic P form in Tshobdun. We propose a historical scenario to explain how such radically different systems came into being, proposing in particular that one of the generic human markers was grammaticalized from a nominalizer, and show that our reconstruction can also explain the origin of the local scenario portmanteau 1→2 and 2→1 prefixes. These reconstructions allow us to establish the existence of several previously unattested grammaticalization pathways.
In this chapter I examine the phonetic origins of voiceless sonorants cross-linguistically within the general framework of Evolutionary Phonology (Blevins 2004; 2006; 2008b; 2015). In terms of a general hierarchy of contrast, we observe that: voiceless obstruents are common; voiceless sonorant consonants are uncommon; voiceless vowels are extremely rare. One phonetic source of voiceless sonorants is coarticulation in RH and HR and clusters, where R is a sonorant and H is a segment produced with a spread glottal gesture. Voiceless sonorants may also arise when laryngeal spreading gestures are associated with prosodic domains. In this second case, voiceless sonorants can arise as allophones of their voiced counterparts. While a fair number of languages show voiceless sonorant glides, liquids and nasals phonologized as a consequence of RH/HR coarticulation, voiceless vowels resist phonologization despite their high frequency as phonetic variants of modal vowels. In some cases, voiceless vowels are lost before phonologization can occur. In other cases, resistance to phonologization may be due to effects of analogy, /h/, word phonotactics, or lexical competition.
The Obligatory Coding Principle accounts for the inventories of possible coding frames in languages that, according to the current terminology, can be characterized as consistently accusative or consistently ergative in their system of argument coding. In coding frame inventories fully consistent with the Obligatory Coding principle, every coding frame includes a given type of coding, either A (in obligatory A coding languages) or P (in obligatory P coding languages). However, languages with coding frame inventories violating this principle are not exceptional. This chapter examines the types of evolutions that may result either in global shifts affecting the Obligatory Coding Principle, in systematic violations of the Obligatory Coding Principle, or in the gradual spreading of non-canonical coding frames. The idea underlying this study is that, before discussing the theoretical status of this kind of generalization, it is crucial to clarify its involvement in diachronic processes.
A central issue in typology is the role of implicational hierarchies in shaping individual languages. One view is that the hierarchies guide language change, or at least constrain it: “Since a hierarchy constrains what is a possible language, it is also a constraint on language change, because languages move from one possible state to another” (Corbett 2011). Other approaches take a different perspective: “Hierarchies simply capture the outputs of independent diachronic processes” (Cristofaro & Zúñiga this volume). Here the relationship between typology and diachrony is examined with respect to the most frequently-cited hierarchies, the cluster of Referential/Topicality/Animacy/Empathy hierarchies. While such hierarchies might appear to drive diachronic development in some single-step changes, multi-step developments are a different matter.
Sahaptin and Nez Perce, the two languages of the Sahaptian family, have both been cited as case studies in the typological literature on hierarchical patterns in main clause grammar. Nez Perce has ergative case marking on only third person singular transitive subjects, plus a minor pattern of indexation of SAP participants via (rarely occurring) second position enclitics. Sahaptin has one of the more complex hierarchical systems ever described, with SAP indexation via enclitics, third person indexation on verbs, differential object marking, an inverse verbal direction prefix, and two distinct ergative suffixes, each restricted to a subset of third person singular transitive subjects (one when objects are SAP, the other when objects are third person). This paper begins by reviewing, evaluating, and occasionally expanding on existing knowledge: we summarize the hierarchical patterns in Sahaptian and characterize each distinct construction. Then we compare relevant Sahaptin morphemes with cognates in Nez Perce, and review their reconstruction to Proto-Sahaptian. The primary contribution of this paper is organizing the morphemes (and their accompanying hierarchical patterns) in both languages into cognate constructions, then reconstructing each to its Proto-Sahaptian origins. We conclude by reviewing and evaluating proposals for Pre-Proto-Sahaptian developments claimed to explain the origins of hierarchical patterns that reconstruct to Proto-Sahaptian. The mechanisms we identify as having created the Sahaptian hierarchical effects are diverse, some motivated and others not, some arising from internal sources, others arguably from contact.
A set of Old Irish clitic person markers may co-index subject or object markers on the verb. Only one clitic of this set may appear at a time, however. The clitic, if present, indexes the verbal argument highest on the scale 1st person > 2nd person > 3rd person animate > inanimate. The restriction of one clitic per verb, a prerequisite for the hierarchy, is explained here as a result of the deictic origin of the clitic set. The deictic origin further favors the marking of local persons over non-local persons. For the upper and lower end of the hierarchy, however, the ranking appears to be determined rather by considerations of function. Since the clitics can mark topics, they favor 1st persons over 2nd and animates over inanimates.
This paper reconstructs the history of a set of innovated 1st and 2nd person verbal prefixes in Reyesano which manifest the phenomenon of ‘hierarchical agreement’ in transitive clauses, according to a 2>1>3 hierarchy. I argue that these prefixes come from independent ergative-absolutive pronouns which first became case-neutral enclitics in 2nd position in main clauses and then verb prefixes. And I show that the hierarchical effects that the prefixes manifest in synchrony have nothing to do with the working of a hierarchy during the grammaticalization process. In doing so, the paper contributes to the growing body of diachronic evidence against the idea that the person hierarchy is a universal of human language reflecting a more general principal of human cognition.
In this chapter, we extract general principles of language change from the study of the evolution of the conjunct order in various Algonquian languages, and propose four generalizations concerning the directionality of the spread of analogy in these systems. These generalizations are expected to bring insights on the analysis of data from other language families with direct/inverse marking but insufficient philological record, such as for instance Sino-Tibetan.
Tupi-Guarani languages are supposedly perfect examples of hierarchical indexing systems, where the relative ranking of A and P on the 1 > 2 > 3 person hierarchy determines the selection of the person markers. This chapter questions the relevance of the person hierarchy as a synchronic and diachronic explanation for such systems, with data from 28 languages. First, only SAP > 3 can really be posited in the actual languages, and second, it explains only part of the facts that it is supposed to account for in Proto-Tupi-Guarani. The chapter ends by suggesting that these systems do not result from the person hierarchy as a functional motivation. Instead, they may result from grammaticalization of pronominal paradigms lacking third-person forms.
There are three distinct transitive constructions in four Coast Salish languages, Squamish, Halkomelem, Klallam and Lushootseed. In the V-tr construction, both A and P are unmarked for case; in the V-mid construction (often considered antipassive), A is unmarked and P marked; in the V-tr-mid construction (often considered passive), P is unmarked and A marked. Individually, none of these constructions is a hierarchical system, but in combination, asymmetries in their distribution are well on the way to creating a person-based hierarchical system. This paper discusses the diachronic development of each of these constructions, then describes their differential distribution into the four functional domains: local (SAP A → SAP P), direct (SAP → 3P), nonlocal (3A → 3P), and inverse (3A → SAP P). While the distribution is not identical in each of the languages, the trend is clear: the etymologically passive V-tr-mid construction cannot occur in the direct domain and has become the pragmatically unmarked construction in the inverse domain, whereas the etymologically antipassive V-mid construction cannot occur in the inverse domain. While it only occurs in the direct and nonlocal domains, even there it is rare, giving the appearance that its function is still that of an antipassive. In combination, the result is that whenever the two core arguments of a clause are an SAP and a third person, regardless of grammatical role the SAP participant is always an unmarked core argument, whereas the third person is most often marked, leading to a situation where the oblique case in these languages is beginning to resemble the obviative case-marker of inverse languages.
The study of hierarchical argument indexation systems shows that while the ranking of both 1st and 2nd person over other arguments is robust and reliable, it is impossible to find any compelling crosslinguistic evidence for one or the other ranking of the two Speech Act Participants, and rare to find a consistent ranking even within a single language. This paper assembles and reviews historical changes in the indexation of the “local” categories (1→2 and 2→1) in a number of Tibeto-Burman languages. We see that the fundamental deictic ranking SAP > 3 is conservative, and inverse marking to emphasize that ranking has been reinvented several times in the family. Changes in the marking of local categories are more diverse, but two phenomena recur independently in different languages and branches: a tendency for the 1→2 form to be uniquely marked, sometimes with forms which are not synchronically relatable to anything else in the paradigm, and a contrasting tendency for the 2→1 form to merge with the marking of 3→1. I propose that these tendencies reflect what I call sociopragmatic effects, i.e. the socially delicate nature of any and all natural utterances involving both the speaker and the addressee.
The Animacy Hierarchy (AH) is an important construct employed for the description and explanation of variation and splits in case marking and agreement in various grammatical domains. The AH is a scale that combines person, definiteness and semantic animacy and is used to state clear preferences of certain morphosyntactic coding types over others. One assumption of the AH is that proper names (PNs) occupy an intermediate place between personal pronouns and common nouns. Despite the large body of research since its first extensive formulation in Silverstein (1976), it is astonishing that there has been almost no empirical evidence published for this claim. Since the AH has been formulated mostly on the basis of case marking and agreement phenomena in languages with split ergativity or hierarchical alignment, we compiled a sample of more than 30 such languages in order to find data on the morphosyntactic coding of PNs. While there are only a very few instances that confirm the claim, there are more instances that contradict it. We concluded that PNs should be removed from the AH, since their assumed position has no predictive value for typological generalizations.
This paper discusses the history of generic person marking systems in several Gyalrong languages. While closely related, Japhug and Tshobdun differ considerably: the inverse prefix marks generic A in Japhug, while it appears in the generic P form in Tshobdun. We propose a historical scenario to explain how such radically different systems came into being, proposing in particular that one of the generic human markers was grammaticalized from a nominalizer, and show that our reconstruction can also explain the origin of the local scenario portmanteau 1→2 and 2→1 prefixes. These reconstructions allow us to establish the existence of several previously unattested grammaticalization pathways.