This feature is inspired by WALS feature 129 (Brown 2005) and concerns the semantic identity or differentiation between the words meaning ‘hand’ and ‘arm’. There are languages that have different words for forearm and upper arm (like Gurindji, one of Gurindji Kriol’s lexifiers, which has wartan ‘hand and forearm’ and murlku ‘upper arm’; Gurindji Kriol), but for the purposes of this feature, we consider only the forearm, from the elbow downwards. Of the several logical possibilities to partition the semantic space of the upper limb, Brown (2005) maps two, (a) identity, where one word means both ‘hand’ and ‘arm’ and (b) differentiation, where there are two words, one denoting ‘hand’ and another ‘arm’. Brown’s identity subsumes cases where there are two different (possibly related) words, but one of them denotes ‘hand and arm’ and the other only ‘hand’ or only ‘arm’, but we assign this constellation to a different value, overlap. A fourth constellation, identity and differentiation, where there are at least three words—one denotes ‘hand’, one denotes ‘arm’, and the third denotes ‘hand and arm’—is not attested among the APiCS languages.
Three strategies to refer to hand and/or arm are found among the contact languages in APiCS:
Value 1. ‘Hand/arm’ identity is the most frequent value in our sample, 43 languages (57%) showing this strategy. All APiCS languages with non-European lexifiers show ‘hand and arm’ identity, e.g.
There are also some contact languages with a European lexifier that have ‘hand and arm’ identity:
Value 2. Differentiation, that is, two different words for ‘hand’ and ‘arm’, is found in 27 APiCS languages (36%), all of them having a European lexifier. Compare:
Value 3. Only five of the languages sampled in APiCS (7%) show semantic overlap, i.e. two different (possibly related) words, of which one denotes ‘hand and arm’ and the other only ‘hand’ or only ‘arm’. Again, all of these languages have a European lexifier: Bahamian Creole, Cape Verdean Creole of São Vicente, Chinese Pidgin English, Ghanaian Pidgin English, and Sri Lanka Portuguese. In all cases, it is the lexifier word designating the hand which means ‘hand and arm’ in the contact language, while the lexifier word designating the arm means ‘arm’ only. Compare Cape Verdean Creole of São Vicente, where mon (< Portuguese mão ‘hand’) refers to the hand and the arm, while bros (< Portuguese braço ‘arm’) refers to the arm only:
Differentiation is the most frequent value in Brown (2005), with 389 of 617 languages (63%) following this pattern. By contrast, 48 of 75 APiCS languages (64%) show ‘hand and arm’ identity (in value 1 and value 3). This is all the more remarkable in view of the facts that (i) 61 of 76 APiCS languages have a primary European lexifier and (ii) all European lexifiers in APiCS except Russian differentiate between ‘hand’ and ‘arm’. The explanation may lie in Brown’s (2005:522) observation that there are more identity languages closer to the equator: Most substrates of APiCS languages derive from these latitudes and the overall substrate pull would have been towards ‘hand and arm’ identity.
With only one exception (Berbice Dutch, where bara comes from the substrate Ijo bárá ‘hand and arm’), APiCS languages derive the words designating parts of the upper limb from their lexifier languages. Thus, value 1 languages either inherited the ‘hand and (fore)arm’ identity from the lexifier or generalized one word from a ‘hand’ vs. ‘arm’ differentiating lexifier to cover both ‘hand and (fore)arm’. Identity appears to be inherited from the lexifier, among several others, in Ambon Malay, Singapore Bazaar Malay, and Sri Lankan Malay, which derive their words from Malay tangan ‘hand, forearm’, or Chinese Pidgin Russian, whose ruka ‘hand, arm’ mirrors the semantics of the Russian etymon ruka. Interestingly, if the lexifier differentiates between ‘hand’ and ‘arm’ but the contact language shows ‘hand and arm’ identity, it was the lexifier word for ‘hand’—and not that for ‘arm’—that was generalized to mean ‘hand and arm’. Presumably this is because the hand is the most salient part of the upper limb. This happened both in languages with a European lexifier (e.g. Bislama han ‘hand, arm’ < English hand, Diu Indo-Portuguese mãw ‘hand, arm’ < Portuguese mão ‘hand’, or Media Lengua manu ‘hand, arm’ < Spanish mano ‘hand’) as well as in those with non-European lexifiers (e.g. Fanakalo sandla ‘hand, arm’ < Zulu isandla ‘hand’ or Gurindji Kriol wartan < Gurindji wartan ‘hand and forearm’ as opposed to Gurindji murlku ‘upper arm’). Yimas-Arafundi Pidgin, which may have generalized Yimas maŋkaŋ ‘arm’ rather than nuŋkara ‘hand’, possibly constitutes the only exception to this rule, but Foley (2013) is “uncertain” about the value 1 status of this language.
Similarly, with regard to value 3 (overlap) it is the lexifier word for ‘hand’ that is widened to mean ‘hand and arm’ in the contact language, while the word for ‘arm’ does not undergo any semantic change. Compare example (7), above, or Sri Lanka Portuguese:
The semantic widening of lexifier ‘hand’ to contact language ‘hand and arm’ (as covered by value 1 “identity” and value 3 “overlap”) is found to varying degrees in English-lexifier (21 of 27 English-lexifier APiCS languages), Portuguese-lexifier (10/14), and Spanish-lexifier (1/5) APiCS languages. However, not a single of the ten French-lexifier contact languages in APiCS abandoned the main ‘hand’ vs. bras ‘arm’ differentiation of French, even though there must have been a considerable substrate pressure at least in the Caribbean and on New Caledonia (most West and Central African languages as well as most Melanesian languages showing ‘hand and arm’ identity; Brown 2005: 524-525).