Chapter 115: 'Hear' and 'smell'

Feature information for this chapter can be found in feature 115.

1. Feature description

Map 115 shows the identity or differentiation of the words expressing the meanings ‘to hear’ and ‘to smell’, the latter used transitively.

In many languages of the world, one or several sense perceptions and/or processes of cognition are expressed by the same verb (cf. Evans & Wilkins 2000, Vanhove 2008). Compare e.g. French entendre, which means both ‘to hear’ and ‘to understand’, or sentir, which can mean ‘to feel’, ‘to smell’, ‘to taste’ or more generally ‘to perceive’. Such formal identity in the semantic domains of sense perception and cognition is also common among the APiCS languages. For example, Bislama harem and Kinubi ásma both mean ‘to hear’ and ‘to feel’, while Tayo sa:(ti) and younger generation Principense xintxi mean ‘to smell’ and ‘to feel’. A very common case of identity is that between ‘to hear’ and its metaphorical extension ‘to understand’, as found, among others, in Eskimo Pidgin tusar-, Ghanaian Pidgin English hiɛ, Guinea-Bissau Kriyol obi, Palenquero kuchá or Sango ma. Some contact languages have a general verb of sense perception, like Nigerian Pidgin, where “hyar ‘hear’ can be used to refer to stimuli detected by any of the senses” (Faraclas 2013).

The present feature explores the extent of overlap in the area of sense perception by focussing on the words meaning ‘to hear’ and ‘to smell’.

2. The values

This feature has four values:

Differentiation63
Identity7
Overlap3
Identity and differentiation1
Representation:74

Value 1. We speak of differentiation if one word denotes ‘to hear’ and another word denotes ‘to smell’. Note that this value also covers cases of (accidental) phonological overlap, as illustrated by Pidgin Hindustani suno ‘to hear’ vs. suŋo ‘to smell’, because they constitute separate words even though they are partially identical.

Differentiation of ‘to hear’ and ‘to smell’ is by far the most common value in APiCS: 85% of the contact languages in our sample use different words to encode ‘to hear’ and ‘to smell’, as illustrated by Norf’k:

(1)

ya ‘to hear’
smael ‘to smell’
(Mühlhäusler 2013)

Value 2: There is identity if a single word is used to express ‘to hear’ or ‘to smell’ and no word exists that denotes only ‘to hear’ or only ‘to smell’.

This is the case in Fanakalo, where izwa means both ‘to hear’ and ‘to smell’ and where there is no other word whose meaning is restricted to either ‘to hear’ or ‘to smell’. A number of APiCS contributors point out that the verb used has an even more general meaning, covering more sense perceptions than just ‘to hear’ and ‘to smell’. In Kikongo-Kituba, for example, “the basic meaning of wa is ‘perceive’” (Mufwene 2013).

Because of the semantic indeterminacy of the verb, the context plays a crucial disambiguating role. The particular interpretations of ‘to hear’ or ‘to smell’ arise in the context and/or are created only in conjunction with the other constituents that the general verb of sense perception combines with. Most often this is the object expressing the perceived entity, as in

(2)
a.
Mono
1sg
mene
prf
wa
perceive
makelele.
noise
I have heard noise.
b.
Mono
1sg
mene
prf
wa
perceive
nsudi.
smell
I have smelt (sth.).

In one APiCS language, however, the disambiguation is achieved not by the object of perception but by a prepositional phrase referring to the organ of sense perception:

(3)
a.
kámtaks
know
kápa
prep
q'walên
ear
to hear
b.
kámtaks
know
kápa
prep
nus
nose
to smell

It could be argued that the constructions illustrated by (2) and (3) actually present cases of differentiation rather than identity since in addition to the (identical) verb other (different) sentence constituents are necessary to generate the semantic concepts of ‘to hear’ and ‘to smell’. Following this argumentation, the particularized meaning ‘to hear’ in Kikongo-Kituba is represented by wa makele and ‘to smell’ by wa nsudi. Seen this way, while there is formal overlap between the constructions expressing the two sense perceptions (wa), we are dealing with essentially different forms because the two constructions are not identical (makele vs. nsudi). However, this line of argumentation presupposes that the constituents form part of the verb, which they do not, or that wa makele and wa nsudi are fixed expressions, for which there is no evidence. Also, perception verbs are explicitly or implicitly transitive and thus usually require an object, no matter whether or not a language has an abstract perception verb (‘to perceive’) or particularized perception verbs (‘to hear’ and ‘to smell’). Since this object is necessarily different for hearing and smelling, both in languages assigned to value 1 (differentiation) and to value 2 (identity), they cannot form the basis of classification. This leaves us with the verbs only, which seems to correspond to the intuition of the speakers of “identity” language: At least for (Mufwene, p.c.) the contributor’s “native speaker intuition is that the same form is being used in all these constructions, pace the English translation” Kikongo-Kituba.

Identity between ‘to hear’ and ‘to smell’ is reported for earlier stages of some APiCS languages. For example, Bislama harem and the related Tok Pisin harim covered both meanings in earlier times (Meyerhoff 2013; Smith & Siegel 2013). It seems that this identity was broken up through pressure from the lexifier English, whose smell was borrowed to cover the olfactory perception. As a consequence, the meaning of harem/harim was narrowed to auditory perception, just as in English.

Value 3. Overlap refers to cases where there are two different (possibly related) words, but one of them denotes ‘to hear’ and ‘to smell’, and the other one denotes only ‘to hear’ or only ‘to smell’.

There are only three languages in the sample where this is the case, all belonging to the West African KrioPidgin English continuum. In Krio, Ghanaian Pidgin English and Nigerian Pidgin the word derived from English hear means ‘to hear’ and ‘to smell’, while the word derived from smell means ‘to smell’ only. This overlap is probably an intermediary stage between identity and differentiation, in which an older word covering both hearing and smelling (identity) coexists with the newly borrowed English smell (incipient differentiation). For full differentiation to develop, the more general older word will have to undergo semantic narrowing to refer to auditory perception only.

Value 4. The identity and differentiation constellation is a combination of values 1 and 2: There are at least three (possibly related) words; one denotes only ‘to hear’ and another denotes only ‘to smell’ (differentiation), and the third denotes ‘to hear’ or ‘to smell’ (identity).

Casamancese Creole (Biagui & Quint 2013) is the only language in the APiCS sample that fulfils these criteria: Wobí means ‘to hear’ (4a) and kerá ‘to smell’ (4b), while sintí, with a generalized meaning ‘to feel’, is used for both hearing (5a) and smelling (5b):

(4)
Differentiation: wobí 'to hear' vs. kerá 'to smell'
a.
N
1sg.sbj
Ø
pfv
wobí
hear
añju
child
na
prog
corá
cry
fora.
outside
I hear the child crying outside.
b.
I
3sg.sbj
Ø
pfv
kerá
smell[verb]
keru
smell[noun]
di
of
tabaku.
tobacco
S/he smelt a/the smell of tobacco.
(5)
Identity: sintí 'to hear' or 'to smell'
a.
N
1sg.sbj
ka
neg
Ø
pfv
sintí
feel
wora
hour
ku
rel.obj
mininu
child
Ø
pfv
kordá.
wake.up
I did not hear when the child woke up.
b.
Mininu
child
ta
hab
sintí
feel
keru
smell
di
of
si
poss.3sg
mamɛ.
mother
A child is able to recognize (smell) its mother's smell.

3. Distribution

Values 2 (identity), 3 (overlap) and 4 (identity and differentiation) all involve identity, that is, a word that covers both auditory and olfactory perception. In this respect, these values contrast with value 1 (differentiation), where there is no word that both means ‘to hear’ and ‘to smell’. The map shows that with one exception (Chinuk Wawa) all value 2, 3 and 4 languages are located in Africa or on the Gulf of Guinea islands. Since these languages have diverse lexifiers – African (Fanakalo, Kikongo-Kituba, Lingala, Sango), English (Ghanaian Pidgin English, Krio, Nigerian Pidgin) and Portuguese (Angolar, Casamancese Creole, Principense) – and English and Portuguese do not have a hearsmell identity, this suggests that it must be due to African substrate and/or adstrate influence. Indeed, Viberg (1984: 141–142) and Welmers (1973: 476) report a wide distribution of ‘hear’– ‘smell’ identity in sub-Saharan languages. However, the absence of identity in the African diaspora creoles on the other side of the Atlantic remains a puzzle. More research is clearly needed, including a comprehensive world-wide typological study on perception verbs and an analysis of perception verbs in earlier stages of contact languages.