by D.N.S. Bhat
Map 43A examines the relationship between third person pronouns and demonstratives on the basis of data gathered from 225 languages. The primary division here is between languages in which third person pronouns and demonstratives are unrelated and languages in which they are either identical or derivationally related. The latter languages are further divided into five different groups depending upon the type of relation that exists between the two.
|Third person pronouns and demonstratives are unrelated to demonstratives||100|
|Third person pronouns and demonstratives are related to all demonstratives||52|
|Third person pronouns and demonstratives are related to remote demonstratives||18|
|Third person pronouns and demonstratives are related to non-remote demonstratives||14|
|Third person pronouns and demonstratives are related by gender markers||24|
|Third person pronouns and demonstratives are related for non-human reference||17|
Among the 225 languages examined, 100 belong to the first group, in which third person pronouns are unrelated to demonstratives. In Ainu (Hokkaido, Japan), for example, the third person pronoun is sinuma, which is unrelated to the demonstratives tan ‘this’ and toan ‘that’ (Tamura 2000: 48, 93). Similarly in Epena Pedee (Choco; Colombia), iru ‘third person’ is unrelated to the demonstratives na ‘this’ and hã ‘that’ (Harms 1994: 45, 57).
The remaining 125 languages show either identity or derivational relationship between third person pronouns and demonstrative pronouns. These languages are divided into five different types (values 2-6), depending upon the type of affinity that obtains between the two.
In the case of 52 of these 125 languages, the affinity derives either from the fact that any of the demonstratives can be used as third person pronouns, or from the fact that third person pronouns and demonstratives share the same stem. In Basque, for example, hau ‘proximal’, hori ‘medial’ and hura ‘distal’ are third person pronouns that also function as demonstrative pronouns (Saltarelli et al. 1988: 213). 33 languages (out of the 52) have demonstratives any one of which can be used as a third person pronoun.
In the case of four languages, there is a distinct third person pronoun, but the demonstratives can also function as third person pronouns. In Yukulta (Tangkic; Queensland, Australia), for example, ṇiya is the third person pronoun, but the demonstratives, ṭanta ‘this one’ and ṭant̪inta ‘that one’, can also function as third person pronouns (Keen 1983: 213).
The remaining languages that belong to this group show a derivational relationship between third person pronouns and demonstratives. This involves the use of the same stem in the case of seven languages, and the use of the third person pronoun itself for deriving demonstrative pronouns (or vice versa) in the case of eight languages.
In Axininca Campa (Arawakan; central Peru), for example, the third person pronouns irirori ‘he’ and iroori ‘she’ on the one hand, and the demonstratives irika (M), iroka (F) ‘proximal’, irinta (M), ironta (F) ‘medial’ and irintó (M), irontó (F) ‘distal’ on the other, all share the stem ir- (Reed and Payne 1986: 324, 330).
In Ambulas (Sepik; Papua New Guinea), on the other hand, demonstrative pronouns are derived by attaching the deictic markers kén ‘proximal’ and wan ‘remote’ to third person pronouns (Wilson 1980: 53, 55). This is also true of Khasi (Mon-Khmer; northeastern India), in which the demonstratives are derived by adding the third person pronouns u (M) and ka (F) to – ne ‘proximal’, - to ‘medial’ and – tay ‘remote’ (Nagaraja 1985: 10-11).
Kharia (Munda; central India) shows the reverse type of relationship in that the third person pronouns are derived by attaching kar ‘person’ to the demonstratives u ‘this’, ho ‘that’, han ‘remote’ and hin ‘invisible’ (Biligiri 1965: 37).
The affinity between third person pronouns and demonstratives is restricted to remote demonstratives in the case of 18 languages. Again, the affinity can consist in exact identity or in a shared stem. In the case of 9 of these 18 languages, third person pronouns are identical with remote demonstratives. In Eastern Armenian, for example, the third person pronoun is na, which is identical to the remote demonstrative na, but distinct from the proximal demonstrative sa and medial demonstrative da (Kozintseva 1995: 12-13).
On the other hand, there are nine languages in which the third person pronouns share their stem with the remote demonstratives. For example, in Lower Grand Valley Dani (Trans-New Guinea; southern Papua, Indonesia), at ‘third person’ shares a stem with the remote demonstrative aty ‘that’, which is distinct from the proximal demonstrative jy ‘this’ (Bromley 1981: 190, 207).
5 of these 9 languages show an affinity of this type only between the non-singular form of the third person pronoun and the remote demonstrative. For example, Hawaiian (Polynesian) has the third person pronoun ia, which assumes the stem lā in the dual and plural, and the distal demonstrative kēlā (which has the alternants lā and ana) (Elbert and Pukui 1979: 107, 110).
The affinity is restricted to non-remote demonstratives in the case of 14 languages. The affinity is with proximal demonstratives in the case of four of these languages. In Asmat (Trans-New Guinea; southern Papua, Indonesia), for example, the third person pronoun is ar, which also functions as the proximal (‘near 1’) demonstrative, contrasting with the two non-proximal pronouns ja ‘near 2’ and er ‘distal’ (Voorhoeve 1965b: 142, 155).
In the case of the remaining 10 languages, the affinity is between third person pronouns and medial demonstratives. Brahui (Dravidian; Pakistan), for example, has the demonstratives dā ‘proximal’, o ‘medial’ and ē ‘distal’, of which the medial one also functions as third person pronoun (Bray 1986: 84).
In the case of 8 of these languages, the demonstrative that is related to the third person pronoun is not specified as “medial”, but it does nevertheless contrast with a distinct ‘distal (yonder)’ demonstrative. For example, Lezgian (Nakh-Daghestanian; eastern Caucasus) has the demonstratives im ‘proximal’, am ‘distal’ and at’a ‘remote (yonder)’, of which am functions as third person pronoun (Haspelmath 1993: 184-188).
In the case of 24 languages, the affinity derives from the recurrence of the same set of gender (or noun class) markers. For example, Apurinã (Arawakan; northwestern Brazil) has the third person pronouns u-wa ‘he’ and o-wa ‘she’, and demonstrative pronouns u-kira (M), o-kira (F) ‘distal’ and i-ye (M), o-ye ‘proximal’ (Facundes 2000: 350, 361). Similarly, the third person pronouns of Venda (Bantu; South Africa) show a distinction for 21 noun classes, and each of the language’s four demonstrative pronouns also makes a similar distinction for the 21 noun classes (Poulos 1990: 95, 106).
9 of these languages are described as showing a masculine-feminine gender distinction, and three as showing a masculine-feminine-neuter gender distinction; 12 of them, on the other hand, are described as showing different kinds of noun class distinctions. In the case of one of these languages, Grebo (Kru; West Africa), the distinction is between large, important things (including humans) and small, unimportant things (Innes 1966: 46, 50).
In the case of some of the languages in which third person pronouns are identical with demonstratives (values 2-4), we do find gender distinctions occurring among forms that function as both third person pronouns and demonstratives. For example, Malayalam (Dravidian; Kerala, India) has demonstratives avan (M), avaḷ (F) ‘remote’ and ivan (M), ivaḷ (F) ‘proximal’ that can also function as third person pronouns (Asher and Kumari 1997: 258), and these show a gender distinction. However, such languages are not included in this fifth group. Only languages in which third person pronouns are distinct from demonstratives, but are nonetheless similar in sharing the same set of gender markers, are included in this group.
In the case of the remaining 17 languages, the affinity derives from the fact that third person pronouns are used only for referring to human or animate beings. In order to refer pronominally to other beings or inanimate objects, these languages make use of their demonstrative pronouns. For example, the third person pronoun upa of Jaqaru (Aymaran; Peru) is used for referring to humans only; the demonstratives aka ‘this’ and uka ‘that’ are used for referring to non-humans (Hardman 2000: 27).
Since the notion of affinity or “derivational relationship” is a graded one, the division between the first group (unrelated) and the rest (related) cannot be sharp and clear-cut. It is quite possible that some of the languages that are included in the first group could also be classified as belonging to one of the other groups. For example, Serbian-Croatian has the third person pronouns ôn ‘he’, òna she’ and òno ‘it’, and the demonstrative pronouns ovaj ‘near 1’, taj ‘near 2’ and onaj ‘distal’ (Kordić 1997: 22-25). It is possible that one or more of these pronouns (especially onaj ‘distal’) share the stem with the third person pronouns. Similarly, Mundari (Munda; northwestern India) has the third person pronoun ay, which appears to resemble the demonstrative pronouns niy ‘proximal’, iniy ‘medial’ and hiniy ‘remote’ in its ending (Osada 1992: 64, 68). Eastern Pomo (Hokan; northern California) has the pronouns mi·p’ ‘he’ and mi·ṭ’ ‘she’, which appear to be correlatable with the proximal demonstrative mé ‘this’ (McLendon 1975: 124, 161). However, none of these languages is regarded here as showing any affinity between third person pronouns and demonstratives.
Third person pronouns are considered to be related to demonstratives only in the case of languages in which either (i) the affinity is quite obvious or (ii) the author has suggested that the two are historically related.
Two of the languages of the sample are reported to have no independent third person pronouns as such. They are Cayuvava (isolate; Bolivia) and Wichita (Caddoan; Oklahoma, Kansas, Texas). In Cayuvava, words that translate as pronouns are derived by attaching the direct object personal affixes to certain auxiliary stems. These also function as auxiliary verbs (Key 1967: 34). In Wichita, on the other hand, personally inflected participles of the verb ‘be’ are given as translations of English pronouns (Rood 1976: 10).
However, in Cayuvava, the verbal prefixes for denoting the person and number of the subject show a deictic distinction in the case of the third person (/ɸ/ ~ /ir/ ~ /t/ ~ /r/ ‘he, she, it’ vs. /k/ ‘he, she, it (far away)’). Words that translate as pronouns are derived by attaching object suffixes (rather than subject prefixes) to the auxiliary verb ãre ‘is, exists’ and therefore do not show this deictic distinction; but we do find such a distinction in the case of the forms me arenahi ‘this one’ and me arekehi ‘that one’ (Key 1967: 28, 36). I have therefore included this language under group 2. On the other hand, Wichita has the demonstratives tiʔi ‘this’, ha:ri ‘that’, the latter of which can occur as a third person pronoun (Rood 1976: 10). I have therefore included this language under group 3.
These typological distinctions do not appear to manifest any striking patterns in their geographical distribution. It might perhaps be suggested that languages showing no relation (group 1) are more common (i) in Africa, (ii) in southeast Asia stretching north to Japan and Korea and east into the Pacific, and (iii) in North America, except for California, than they are in other parts of the world.
The question as to whether third person pronouns are related or unrelated to demonstratives appears to form the basis for a typological distinction between two distinct sets of languages. We might call languages in which third person pronouns are related to demonstratives two-person languages, and those in which the two are unrelated three-person languages. Third person pronouns form part of the system of personal pronouns in the latter case but not in the former. This distinction appears to be correlatable with certain other characteristics such as, for example, that
gender distinctions among third person pronouns are displayed more frequently by two-person languages than by three-person languages, and
the deictic system is less frequently person-oriented in two-person languages than in three-person languages.
For example, among the 125 two-person languages that occur in my sample, 43 show a gender distinction, whereas among the 100 three-person languages, only 13 show a gender distinction (among 7 of these latter languages, the distinction occurs in second person as well). This disparity strongly supports the first correlation mentioned above.
There is a similar kind of disparity in the occurrence of person-oriented deictic systems (involving ‘near the addressee’ as a distinct entity) which supports the second correlation given above. Among the 125 two-person languages that occur in the sample, only 14 have a person-oriented deictic system, whereas among the 100 three-person languages, as many as 31 have a person-oriented deictic system (see Bhat 2004).