3.4.3 Developmental Study

McNeill (1997a) reports the preliminary findings of an ongoing study of gesture-speech systems in English-speaking and Spanish-speaking children from age 3 up. The English-speaking subjects are from Chicago, the Spanish from Guadalajara, Mexico.

McNeill and his team compared the manner gestures accompanying English and Spanish utterances in which manner was present lexically, either via the verb or as a subordinate element to the verb. The aim was to check whether the pattern of having the manner gesture with the manner element, as was the case in English, or as a fog (in some other element), as in Spanish, was connected to the presence or absence of a lexical item incorporating manner, or whether it was a trait that responded to the type of language. The following chart, drawn from McNeill (1997a) summarizes the results:

Column one reflects the percentage of synchronizing the manner gesture with the lexical manner item, column three illustrates the percentage of co-occurrence of the manner gesture with some other element (fog items), and column two indicates the co-occurrence with the general-meaning verb of motion go. Column four illustrates the presence in some of the English-speaking children data of onomatopoeic elements which indicated verbally the manner of movement. Some of the rows of some Spanish age groups are empty because the data had not been analyzed yet. The table shows that Spanish places manner gestures mainly not in the manner item but in some other element of the utterance. English speakers do not do this frequently. Instead, they focus their manner gestures on the manner items. They do this very early in development (90% for 3 year-olds). This shows that it is a trait that is acquired soon, and that it is dependent on the language type.

The following two examples, from an English speaking 4 year-old and a Spanish speaking 3 year-old, illustrate the two patterns:

These examples show that children acquire early on the characteristic mode for signaling manner in their language.

It was also seen that English speakers had the possibility of downplaying the importance of the manner verb by synchronizing it not with a manner gesture but with some other type of gesture. McNeill (1997a) argues that this is an adult trait that has to be learned, and that it is not present in Spanish. The following table demonstrates this:

Explicit (not elided) uses of "rolls" and the gerund "rodando"

The chart reflects the results of comparing English and Spanish utterances where the English verb roll and the Spanish gerund rodando are mentioned explicitly. It can be seen that downplaying this verb with a non-manner gesture is done by adults 57% of the time. Children do this less frequently (20-25%). The evidence indicates that it is something that children have to learn.

Results from other languages also seem to confirm that the typological make-up of a language influences the gestures that appear with motion utterances. Özyürek and Kita (personal communication) have compared English with Japanese and Turkish, two verb-framed languages. They have analyzed the renderings of native speakers of two scenes from the same cartoon used by McNeill. The first scene had Sylvester, the cat, and Tweety, the bird, sitting on window sills, one across the street from the other. The cat swings across the street on a rope to catch the bird. Neither Japanese nor Turkish have a lexical item that, like swing, could express an agentive change of location with an arc trajectory. Turkish and Japanese speakers produced non-arc gestures to represent the cat´s motion, in contrast with the English speaking subjects, who produced no non-arc gestures. The second scene finds Sylvester, who had swallowed a bowling ball, rolling down the hill. For this scene, both Japanese and Turkish require two verbs ‘go rolling’ or ‘descend rolling’ (just like Spanish, ir rodando, bajar rodando). In their narrations of this scene, Turkish and Japanese speakers were more likely to produce more pure rotation gestures (representing only manner) and more pure trajectory gestures (representing path) than English speakers. These results are interpreted as suggesting that the linguistic encoding of manner and path information affects how it is represented in gestures.

Manner gestures synchronized with: Adults:	Manner items§	"go"†	'Fog' Items*	Onom?
Spanish	42%	17%	42%	--
English	75%	5%	15%‡	--
All children included:	-	-	-	-
Sp 3-12	15%	15%	69%	--
En 3-12	64%	6%	8%	16%
Children broken down by age -- small samples excluded:	-- -	--	--	--
Sp 11-12	25%	--	75%	--
En 10-12	71%	--	14%	14%
Sp 8-9	-	-	-	-
En 8-9	55%	10%	10%	5%‡
Sp 7	17%	--	50%	--
En 6-7	60%	7%	14%	20%
Sp 5
En 4-5	36%	--	9%	55%
Sp 3	-	-	-	-
En 3	90%	10%	0%	--
_	§ "rolls", "rodando"	†including tense variants	* Ground, Figure, Path	?onomatopoeia ‡plus 'other'

_	Manner Gesture	Non-Manner Gesture	No Gesture	No Gesture + Non-Manner
Sp Adults	80%	20%	--
En Adults	43%	57%	--	57%
Sp 11-12	83%	17%	--
En 10-12	50%	23%	27%	50%
Sp 8-9
En 8-9	44%	37%	19%	56%
Sp 7	-	-	-	-
En 6-7	42%	8%	50%	58%
Sp 4	-	-	-	-
En 4-5	42%	25%	33%	58%
Sp 3	-	-	-	-
En 3-4	46%	23%	31%	54%