3.5. Summary and discussion

In the previous sections, an analysis has been carried out of widely diverse data from novels and translations, from elicited narratives, from acquisition studies, and from gesticulations synchronized with speech. There are several convergent ideas in all of these studies that will be summarized in this section.

In all of the studies, it was found that Spanish speakers devote less attention to path than English speakers. The claim is that the typological difference between the two languages is the reason that explains this fact. Since English expresses path mainly via satellites to the verb (particles and prepositional phrases in the case of English), several of these path elements can easily accumulate with just one verb. English speakers take advantage of this fact to present elaborate descriptions of trajectories in directed motion events. In fact, directional particles play a fundamental role in children’s early expression of motion in English. By three years of age they are able to combine these satellites with manner verbs, although there are no instances yet of both source and goal in a single clause. This is an ability they will have to develop. By age nine, English-speaking children are able to combine verbs encoding the manner and/or the cause of the movement with multiple path elements.

By contrast, Spanish 3 year-olds usually express directed motion events with a bare verb. At age 4 and 5, many of the children seem to go through a U-shaped learning curve where they try to provide more information about paths by redundantly adding a locational adverb that simply repeats the information provided by the path verb (subir arriba, bajar abajo). By age 9, the majority of directed motion sentences are expressed again with a bare verb. Adults and literary authors provide more information about trajectories, but it is still very poor, especially if compared with what is common in English. Very rarely they present more than just one path element per clause. Even in gestures synchronized with language there is a difference in the presentations of path. Spanish speakers present path in their gestures as an undifferentiated whole, whereas English speakers provide detailed information and tend to segment the trajectories they are talking about. There is no a priori linguistic reason that explains why Spanish sentences like the one below with two path elements are rarely attested:

A verb like caer allows for the expression of both a source and a goal. But the fact of the matter is that Spanish-speakers do not readily use the two together. A possible explanation lies again in the typological distinction. Path verbs include in their meaning one or two of the elements of the path and are geared towards expressing usually only one of the path components, usually the goal, because it is less accessible from the context than the source. Thus, Spanish speakers are habituated to fewer path elements in their motion sentences. Manner verbs, on the other hand, do not include in their meaning any information about path. If a path is expressed with a manner verb, it will have to be more fully elaborated.

Instead of focusing on the descriptions of the paths, Spanish speakers prefer to give detailed information about the locational setting where the motion takes place so that the hearer can easily infer the particular characteristics of the path. Thus, they would tend to say:

Both sentences use path verbs, caer and salir, without any kind of complementation. However, they give a full description of the locational setting of the motion. The youngest children do not have the cognitive acumen for this type of descriptive elaboration. It is something that they will have to acquire in their development. The strategy English speakers use is just the opposite. They usually leave for inference the locational setting. This can be done easily from the complex paths, which usually mention some of the salient landmarks that constitute such setting.

With regards to manner, the evidence again converges towards difference between the two languages owing to their type. Since the verb does not have to encode the path in English, it usually lexicalizes other notions, with manner figuring as the most prominent. This explains why even the youngest English speakers have a repertoire of manner verbs that they use profusely. Spanish speakers pay far less attention to manner, both in the elicited narratives and in the novels and their translations. Nevertheless, they seem to compensate for this lack of manner through the gestures that accompany their speech. But even in the gestures there is a difference. English speakers synchronize their manner imagery with the manner items present in the utterance. Even the youngest children do this, which indicates that it is a characteristic that is learned early. Even more, children at all ages, but no adults, sometimes incorporate an onomatopoeic manner item with which they synchronize their manner gestures. By contrast, Spanish-speakers tend to diffuse their manner gestures throughout the whole utterance as a kind of fog. Moreover, English speakers have the option of highlighting the manner component of the verb with a manner gesture, or of downplaying it by synchronizing the manner verb with some other kind of gesture. This manner modulation is less frequent in younger children, which shows that it is a developmental trait. Spanish speakers do not carry out this modulation. Even when there is an explicit manner item in the utterance (such as a gerund e.g. rodando), they synchronize it with a non-manner gesture much less frequently than English speakers. This indicates that the manner modulation is something that is characteristic of English probably because it is a satellite-framed language. The verb of a directed motion sentence usually lexicalizes the manner of movement. In many instances, especially with the manner verbs with more general meaning, the verb is just filling the verbal slot. It is natural, then, that English speakers have developed a way to downplay the manner meaning by way of gesture to indicate that manner should not be paid attention to. Further study of other languages of the same type as English would be needed to confirm whether this modulation is a typological trait or not.

English and Spanish have been the languages most deeply studied with regards to this typological distinction, most likely because they were the languages that Talmy gave as prototypical examples for the distinction. Nevertheless, it was seen in the preceding sections that the studies involving other languages confirmed the results found for English and Spanish.

Only one study seems to go against what would be predicted by the typology. David Wilkins has investigated Arrernte, a verb-framed language spoken in Central Australia. He has collected narratives of the frog story from adult speakers of the language and has compared them with the English adult stories. Arrernte’s speakers build more complex paths and devote more attention to the dynamics of the movement than English speakers. Wilkins attributes these surprising findings to the ethnographic observation that Central Australian Aborigines have a special interest in motion paths and journeys that is unique to the area. Thus, it seems as if the cultural factors can override the linguistic constraints imposed by the typology. Further study about the interaction between linguistic and cultural factors is needed.

The implications of all of the findings reviewed in this chapter go well beyond the boundaries of language. What is in question, according to these researchers, is the issue of the relationship between language and cognition, as Berman and Slobin explicitly state:

In particular, Slobin has proposed that in acquiring a language, the child learns particular ways of "thinking for speaking", thinking generated because of the requirements of a linguistic code (Slobin 1987).

"Thinking for speaking" is a weak form of the controversial Sapir-Whorf hypothesis, the idea that the way human beings view reality is strongly influenced by the semantic and grammatical categories of their language. This hypothesis is not favourably looked upon in many linguistic circles, in part because of the popularity of Chomsky’s universalistic claims (see Pinker 1994 for a ferocious critique of the Sapir-Whorf hypothesis). Nevertheless, it is an idea that has always been alive in one form or another under the less loaded label of ‘linguistic relativity’.¹ "Thinking for speaking" is a weak form of this hypothesis because it deals only with the mental processes that are involved in constructing linguistic utterances, without making any claims about the influence of language on other kinds of mental processes.

The "thinking for speaking" hypothesis implies that instead of having a set of pre-linguistic concepts into which children simply map the semantic and grammatical categories of the language they are acquiring, those same grammatical and semantic categories help shape the concepts the child is going to use in speaking. The following quotation explains Slobin’s reasons for postulating such a hypothesis:

The way this ‘thinking for speaking’ comes about is as follows:

This ‘thinking for speaking’ idea is reminiscent of the Vygotskyan notion that once language interacts with thought something new is created that changes human thinking permanently (Vygotsky 1989).

David McNeill in his gesture studies has also made similar claims. He considers gestures an essential part of thinking. The fact that gestures for manner and path motion are different across languages reveals that the speakers of those languages have learned to "think" differently in order to speak. For him, for example, the divergent ways to gesture manner, with the manner gesture focused on the manner lexical item, as in English, or diffused as a manner fog throughout the utterance, as in Spanish, are forms of thinking about manner that the speakers of the two languages internalized when they learned the language. According to McNeill, it is precisely this process of "thinking for speaking" that brings together the two types of mental representation, the linguistic, the words, and the visuospatial, the gestures (McNeill 1998, and McNeill and Duncan 1998).

Choi and Bowerman (1991) have also raised similar issues. They claim that since English- and Korean-speaking children show language specific principles of spatial semantic categorization very early, they do actually construct those spatial semantic categories with the help of the linguistic input. This is in sharp contrast with the commonly accepted idea that children’s early spatial words are mapped to previously existent non-linguistic concepts.

Recently, Slobin has started to investigate issues that go beyond the "thinking for speaking" hypothesis. A preliminary study in which Spanish, Turkish and English speakers were asked to recall and report on mental imagery evoked after reading narrative texts has given as result that English speakers experience more frequent and vivid images of manner of motion than Spanish and Turkish speakers, who seem to be paying much less attention to it. Thus, for example, after reading a passage like ". . . started to walk through the mud and stones of a path . . .", Spanish speakers do not report mental imagery of manner of motion. English speakers, by contrast, report vivid and specific mental images that correspond to English verbs like trudge along, not present in the test sentence. This might reflect that the lack of attention paid to manner by a language like Spanish plays a role in the way speakers conceive of manner of motion in general.

The implications of these philosophical issues go well beyond the scope of this dissertation, and thus their ramifications will not be further explored here. In any case, this chapter has vividly shown the vitality of the research that has sprung from Talmy’s typological observations. As Berman and Slobin convincingly conclude, this research is discovering new avenues into the complexities of human language and thought: