Navigation – Plan du site

The multi-modal representation of motion events in Awetí discourse

Sabine Reiter

Résumé

Recent research focusing on the insights gestures may offer into the mental conceptualization of motion events suggests that a number of factors may influence the way motion events are gestured in different languages. This study aims at showing how in Awetí oral discourse other than purely linguistic modes of representation such as gesture and ideophones are obligatory ‘path’ and optional ‘manner’ components of motion events. With regard to ‘manner’ it will be illustrated by corpus examples that the information state of this motion component in discourse determines whether it will additionally be gestured or not. These features suggest that the co-speech gestures in Awetí motion events have a meaning-conveying function and should be treated as a means of communication rather than as a feature which is produced with a low level of speaker awareness, merely providing insights into mental conceptualizations.

Haut de page

Texte intégral

1. Introduction

1Recent research focusing on the insights gestures may offer into the mental conceptualization of motion events (McNeill & Duncan 2000, Özyürek et al. 2005, Chui 2010) suggests that a number of factors such as the lexical and grammatical encoding of meaning components of the event, structuring of information in discourse or even cultural differences in patterns of movement may influence the way motion events are gestured in different languages. After a summary of the results obtained by different researchers for English, Spanish and Mandarin, the respective parameters will be explored in a corpus-based analysis of the Tupian language Awetí.

1.1 Previous studies

  • 1 I have excluded the ‘continuation of a stationary location’ which Talmy (2007: 70) together with mo (...)

2According to Talmy (2007: 70-71), the prototypical ‘internal’ components of a motion event1 are a ‘figure’, i.e., an object which moves or is located relative to another object constituting the ‘ground’. Further, there is a ‘path’ and a ‘motion’, referring to the presence of motion per se in the event. In addition, a motion event can have an external ‘co-event’ associated with it, having a ‘manner’ or ‘cause’ relation to it.

3Verbs and other elements in different languages encode information on ‘motion’, ‘manner’ and ‘path’ to different proportions: In English ‘motion’ and ‘manner’ are typically encoded in a verb and ‘path’ in an adjunct, as shown in (1).

(1)

The pencil rolled off the table. (Talmy 2007: 71, [4])

  • 2 According to whether the ‘path’ component is encoded in or outside the main verb Talmy (2007) class (...)

4In Spanish, by contrast, an adjunct, often consisting of a gerund form, provides information on the ‘manner’ of motion, while the verb encodes ‘motion’ and ‘path’.2 An example is given in (2):

(2)

Sale volando. (Slobin 1996, cited in Özyürek et al. 2005: 200)

‘(He/ she/ it) exits flying.’

5McNeill & Duncan (2000) in their comparative study of English, Spanish and Mandarin established specific patterns for the interaction of language and gesture with respect to motion events for each of the languages. From these they drew conclusions for a respective conceptualization, which to them depends on the grammatical structure of the sentential predicate in English and Spanish, and, in the case of Chinese, predominantly on the discourse structure. According to their observations ‘manner’ in English is gesturally represented, only if it is in the speaker’s focus, whether there is a co-occurring ‘manner’ verb or not. The ‘manner’ component can further be downplayed by using deictic verbs such as come and go instead of ‘manner’ verbs. In Spanish, by contrast, ‘manner’ is rarely expressed in speech but can be gestured independently, synchronizing with ‘path’ verbs or NPs referring to the ground component of a motion event. In Chinese, which the authors classify as a satellite-framed language similar to English, gestures often do not synchronize with the verb but with a linguistic unit towards the beginning of the utterance which constitutes the topic.

  • 3 21.6% of all motion events contained ground information which was gesturally represented in 26.4% o (...)
  • 4 A prerequisite for this hypothesis is that she assumes gesture to be closer to the conceptualizatio (...)

6Chui (2009), pointing out several shortcomings in McNeill’s and Duncan’s analysis of Chinese, presents a detailed analysis for Mandarin based on 245 instances of motion events from elicited cartoon narrations, of which nearly two thirds were gestured. These 245 instances were analysed with regard to their linguistic and gestural encoding of various motion-event components. Chui divided the motion verbs into various types with regard to their encoding of the components ‘manner’ and ‘path’ and found out that 81.6% encode ‘manner’ to some proportion, while 63.6% are single ‘manner’ verbs. In addition, ‘manner’ information in Mandarin can be provided by adverbial and phrasal expressions. Single ‘path’ verbs are extremely rare. Gestures, by contrast, encode single ‘path’ information (73.2%) considerably more often than single information on ‘figure’ (9.8%), ‘ground’ (9.2%) or ‘manner’ (3.2%). Combined ‘manner and path’ gestures are listed with 4.6% (see Chui 2009: 1770, Table 1). Chui further investigated the information state (given vs. new information) as another parameter which may influence the gestural representation. She showed that this does not play a role for the gestures depicting ‘manner’ and ‘path’ (see Chui 2009, Tables 3 and 5). However, the high proportions of gestures depicting ‘figure’ and ‘ground’ in comparison to their low overall linguistic representation, with participants retrievable from the context, suggest that these components may be constrained by information state, occurring mostly when being introduced as new information.3 From the fact that ‘manner’ is predominantly encoded by linguistic means, i.e., in the lexical verbs, but that gestures depict a ‘path’ component, Chui concludes that the latter must be most salient in the conceptualization of motion events.4 This applies not only to Mandarin, where single ‘path’ gestures are predominant, but also – according to the results of the study by McNeill & Duncan (2000) – to English and Spanish.

  • 5 I have avoided to classify the gestures considered in this study according to the McNeill (1992) cl (...)

7The following analysis of a number of representative examples focuses on the interaction between motion verbs, gestures and ideophones in the South-American indigenous language Awetí. It will be shown that in this language, too, ‘path’ appears to be a predominant feature in the conceptualization of motion. The center of interest in this paper, however, is not so much on cognitive processes which may be revealed by gesture but on the communicative function of gestures and the semiotically related phenomenon of ideophones in Awetí motion events.5

1.2 The Awetí speaker community and their language

8Awetí is spoken by about 180 individuals in the multi-ethnic environment of the Upper Xingu within the confinements of the Parque Indígena do Xingu in Mato Grosso, Brazil. Representing a branch by itself within the Tupian family, Awetí bears typical features of these languages: it is agglutinative, both prefixing and suffixing, predominantly head-marking and head-final (see Gabas 2006: 47).

  • 6 See also http://www.unesco.org/new/en/culture/themes/endangered-languages/language-vitality/

9The Upper Xingu cultural area is inhabited by ten ethnic groups speaking languages of three different families (Arawak, Carib, Tupi) and one isolate. This multi-ethnic society is characterized by its linguistic diversity in spite of a common culture which has developed in the last 300 years after the different ethnic groups migrated to the area. Although the Awetí language has a small number of speakers and most of them are bilinguals to differing degree with the Tupi-Guaranian language Kamaiurá, it is currently spoken by all children growing up in either of the two Awetí villages and is taught and used as a language of instruction in Portuguese-Awetí bilingual education at the village schools. Portuguese is increasingly used as a lingua franca in interethnic communication and in some of the neighboring groups has started to replace the native language, but in the Awetí community it only plays a role in the educational domain. Portuguese is spoken fluently by several younger male community members, whereas older members and women have no or a very restricted competence in this language. For all these reasons Awetí is not to be considered a highly endangered language (see Reiter 2010).6

1.3 Database and methodology

  • 7 See www.mpi.nl/DOBES/projects/aweti.
  • 8 The documentation setting could be optimized, if members of the community themselves recorded the d (...)

10Awetí was extensively documented between 1999 and 2006 by Sebastian Drude and Sabine Reiter in the DoBeS Awetí Language Documentation Project which was financed by the Volkswagen Foundation and hosted by the Free University of Berlin.7 The annotated Awetí DoBeS corpus currently consists of almost 20 hours of audio and more than four hours of video data of different discourse genres. Among the video data consulted for this study there are three longer mythological narratives, two autobiographical narratives, and several explanations of cultural activities and descriptions of everyday tasks. One text is a dialogue between three men. The speakers are male and female members of the speech community, all above the age of 30. The myths are told by two trained story-tellers. None of the texts were recorded in an experimental setting. The speakers were asked to talk about their lives or a subject they were specialists in (e.g., fishing, manioc preparation), and the recording took place in the presence of native interlocutors (not captured by the camera). This was meant to obtain data as naturally as possible. It could, however, not be avoided that some of the speakers were intimidated by the camera and the recording situation.8 Therefore, the possibility cannot be excluded that this factor had an influence on the frequency and quality of their overall gesture production. The two professional story-tellers, on the other hand, seemed to feel quite at ease, possibly due to the fact that they were used to perform orally in front of an audience. Although the recording of gesture had not been a focus during data collection, most of the video recordings proved rich and suitable for a gesture analysis. A critical remark needs to be made about the different discourse genres: While mythological or historical oral narratives are a common Awetí discourse form, verbal explanations and descriptions of cultural practices must be related to the specific situation of the documentation. Transmission of such knowledge more naturally occurs via observation and imitation.

11The analysis is entirely corpus-based. The data considered here consists of the approximately 4 hours of video recordings which have been annotated so far. In order to give a first overview over the distribution of Talmy’s semantic components of motion events among verbs, ideophones and co-occurring gestures, two longer texts of different genres (a mythological narrative and a task description) were chosen from the Awetí corpus as a sample. In these all motion events were counted and categorized.

12The two texts were chosen because they represent different genres, have approximately the same length and both contain a high quantity of motion events. The first text, of 28 minutes, is a myth about how the indigenous people received their first manioc plants from the seagull (kakaja). It is told by one of the two professional story-tellers in the community. In the second text, a task description of 24 minutes, a man talks about fishing practices, before and after regular contact with non-indigenous people, mostly speakers of Portuguese, was established.

13Sequences illustrating the interplay of the components of a motion event were then selected from the raw material for a qualitative analysis. Since my interest had originally been on ideophones which, however, are considerably less frequent than verbs in motion events, the material presented in this qualitative section is somewhat biased. Nevertheless, it provides an interesting insight in the interplay of different modes of representation in oral discourse.

14The sequences were transcribed following the conventions for gesture annotation as presented in Kendon (2004). According to Kendon a ‘gesture unit’ can be divided into three different phases: the ‘preparation’, the ‘stroke’ and the ‘recovery’ phase. The term ‘stroke’ refers to the “phase of the movement excursion closest to its apex” (Kendon 2004: 112) or to the phase where the ‘expression’ of the gesture is accomplished, while the ‘preparation phase’ is the movement leading to the stroke. Together these two phases constitute a ‘gesture phrase’ (Kendon 2004: 115). The individual gesture phases in the examples are represented in an extra-glossing line with “~~” symbolizing the preparation and “**” the stroke. The recovery sets in when the articulator, i.e., the hand, relaxes or is withdrawn. This phase is represented by “-.-“. In examples, where a preparation phase and a stroke cannot be clearly distinguished from each other, these are glossed by mixed symbols, i.e., by “*~*” (preparation/ stroke). A hold during a gesture phrase is indicated by an underlined symbol “**”. The articulators are annotated by “rh” (‘right hand’), “lh” (‘left hand’) or “bh” (‘both hands’) respectively.

  • 9 Other motion events were entirely expressed by gesture, either in combination with an NP as the ‘fi (...)

15The interaction of gesture and speech in the ‘ideophonic’ motion events was contrasted with motion events entirely expressed by verbal means, i.e., by verbs and/or locative postpositional phrases.9 In order to capture the meaning of ideophones and gestures, which is schematic rather than referential, the discourse context was taken into account for their interpretation. The discourse context further served to determine the respective information state of the different components of the motion event.

2. The use of ideophones and gestures in Awetí discourse

  • 10 See, e.g., Nuckolls (1996).

16Ideophones are a common phenomenon in oral literatures throughout the Amazonian lowlands.10 A concise definition of ideophones, which can be cross-linguistically applied, is given by Dingemanse (2011: 1): “Ideophones are marked words that depict sensory imagery.”

17Awetí ideophones can be divided into two types: A first type – illustrated by (3a) – depicts activities; a second, less conventionalized and onomatopoeic type – shown in (3b) – depicts sound events involving humans or animals.

  • 11 In each example the information in brackets after the free translation refers to the source file in (...)

(3)

a.

Pyw,

o-

uwyp.

IDEO.pick.up

3coref-

arrow

‘Hei picked up hisi arrow.’ (tal_mamuti_2_8, 1641.368)11

b.

Atĩw atĩw atĩw atĩw.

IDEO.sneeze

‘He was sneezing.’ (tal_mamuti_4_8, 0948.341)

18The second type does not play a role in motion events and will not be further described here. Ideophones in Awetí are prosodically marked. They can, for example, be uttered with expressive vowel lengthening, a specific rhythm, a crescendo or a decrescendo. They can have an expressive morphology in that they are often reduplicated, or they can be repeated several times. Syntactically they occur as independent units, often directly preceding or following a clause with a semantically similar verbal predicate. Awetí ideophones can represent events or salient features of events by themselves, but they can also function as invariant predicates with dependent object constituents. They regularly occur in Awetí discourse, whereby no difference in frequency could be observed between their usage in narratives told by the two professional story-tellers and the texts produced by the less proficient speakers. However, in story-telling events there seems to be a broader variety of ideophones, which also tend to show more internal variation. A simple explanation for this may be the fact that a narrative text often presents a broader variety of events than a text which is based on a specific subject. On the other hand, ideophones are an important fore-grounding technique in Awetí verbal arts and tend to occur at decisive moments in narratives, so that a professional story-teller, who has gone through a training of several years, can be expected to have a large repertoire at his disposition.

  • 12 Thesegestures have iconic and deictic features to differing degree. As will be shown in the example (...)
  • 13 Since the overall number of ideophones in the video data used for this study is relatively low, thi (...)
  • 14 Note that Kita (1997) only uses the term ”iconic gestures”. However, his descriptions of the gestur (...)
  • 15 See example (10) below.

19In languages which have ideophones in their repertoire these have often been associated with gesture. Earlier studies pointing out a close relation between ideophones and gesture are Samarin’s (1971) review of studies on various Bantu languages or Kunene’s (1978) study on the Bantu language Southern Sotho. Nuckolls (2001: 273) has termed ideophones in Pastaza Quechua of the Bolivian lowlands as “verbal gestures” for having many functions in common with these. In the Awetí video data available for this study it could be observed that ideophones consistently co-occur with gestures.12 They are further fully time-aligned with these in that the gesture stroke always falls on the ideophone.13 Such a noticeable correlation was first described by Kita (1997: 392), when he investigated the interaction between ideophones, spontaneous iconic co-speech gestures and expressive prosody in Japanese adverbial ideophones.14 Kita could further show that prosodic peak, if existent, also fell on the ideophone. For this last observation corroborative evidence can also be found in Awetí.15In order to account for these correlations, Kita (1997) proposed a two-dimensional approach for an adequate semantic analysis of ideophones and other linguistic phenomena with a non-arbitrary form-meaning relationship. According to this approach, the meaning of ideophones is not represented in an ‘analytic dimension’ in a human’s mental lexicon, where semantic partials are organized in a hierarchical structure, but in an ‘affecto-imagistic dimension’ “in which language has direct contact with sensory, motor, and affective information” (Kita 1997: 380) and where this information remains modality-specific. The ‘analytic dimension’ of ordinary words, on the other hand, is de-contextualized in the sense that it structures experience in terms of rational concepts such as agentivity or causativity. Nuckolls (1996), when characterizing the semantics of ideophones in Pastaza Quechua, in a similar fashion draws a general semiotic distinction between such sound-symbolic words and other words in a language. She attributes a schematic structure to ideophones. While words in a proposition are structured by cognitive operations involving logic or abstract thought, schemata are based on bodily experience and “exist […] in a continuous, analog fashion in our understanding” (Johnson 1987: 25). This makes them meaningful as a whole and, according to Nuckolls, may for example explain the syntactic isolation of such words, which do not easily combine with linguistic units building up whole meanings from partials (morphemes).

20Gestures, as stated above, will be investigated in this paper in terms of their communicative function. In Kendon’s (2004) approach gestures are as much part of the utterance as verbal expressions in making their own contributions to the overall meaning. He lists several different functions of representational (i.e., iconic) co-speech gestures (see Kendon 2004: 161, 175):

  1. They provide a parallel expression to the meaning which is conveyed verbally.

  2. They refine, qualify, or restrict the meaning conveyed by words.

  3. They provide aspects of reference which are not present in the verbal components, e.g., spatial and orientational information.

  4. They create an image of an object that is the topic of the spoken component.

  5. They provide a dimension of visual animation to an account to convey a much richer sensory experience to the interlocutor.

21Deictic (or pointing) gestures, by contrast, may add to the utterance meaning by pointing to a concrete or abstract object referred to in the discourse (Kendon 2004: 160).

22All these functions can be observed in the Awetí examples of motion events to be discussed. In combination with Awetí ideophones gestures have two semantic functions. They narrow down the interpretation of an ideophone by visualizing the same perceptually salient aspect(s) of the activity. In the specific case of locomotion events, however, gestures usually express a different meaning component than ideophones, i.e. the “path” component, as will be shown in the following section.

3. Motion events in Awetí

23Motion events in Awetí, which can be expressed by finite verbs, postpositional phrases, certain hortative expressions (e.g., nawỹj ‘let’s go!’), and ideophones are, with some regularity, accompanied by gesture. Gestures in motion events, we argue, not only constitute a parallel expression to words but can also take over meaning components by themselves. Similar observations hold for ideophones, so that motion events in Awetí discourse can be characterized as being expressed in various semiotic modes. Corroborative evidence for this empirical observation will be given by examples from the Awetí corpus, whereby most of these will be from the two sample texts.

  • 16 These are the two most widespread types of motion verbs according to the typology established by Ta (...)
  • 17 Different from their English translation equivalents go and come which are classified as deictic ve (...)

24Awetí motion verbs can be divided into two classes according to the semantic components they lexicalize: those of the first class contain information on ‘motion’ and ‘path’, while the second class encodes the components of ‘motion’ and ‘manner’.16 Examples for the first class are to (‘go (away)’), ut (‘come’), or tem (‘go out’).17 The second class of verbs is considerably less frequent in discourse. It is represented by verbs such as tan (‘run’), eko (‘walk’), or wure (‘fly’). Combinations of types, e.g., ‘manner-path-deictic verbs’, as stated by Chui (2009) for Mandarin, could not be observed. Both types of verbs may or may not occur with a postpositional phrase indicating ‘ground’ (e.g., nãpeputywã ‘on his former path’), ‘goal’ (e.g., tam ywo ‘to/at the village’), ‘source’ (e.g., nãti ‘from there’) or ‘manner’ of the movement (e.g., pisicleta ‘apo ‘by bike’) or with a subordinate verb in the gerund adding a ‘manner’ component (e.g., Otan ti otwaw a’yn ‘He went away running’). Taking into account the high frequency of the first type of verbs in the data, it can be said that Awetí speakers predominantly use the verb-framed strategy.

25Ideophones in Awetí motion events are similar to the second group in that they also provide information on motion and on the manner of the movement. The latter includes factors such as suddenness, rate, force-dynamics, or terrain. Examples are tutututu for ‘running’ as opposed to tyk tyk tyk for ‘(person) walking’, popopo for ‘flying’, tirik tirik for ‘(person) walking on dry leaves’, or tom for ‘(heavy object/person) falling/diving into the water’. Ideophones depicting motion most often occur as independent clauses adjacent to a clause with a finite verbal predicate referring to the same event.

  • 18 Further Awetí video data needs to be fully annotated in order to accumulate a significant number of (...)

26Gestures in motion events depict the ‘motion’ and the ‘path’ component and may additionally express ‘manner’. Gestures occurring with motion verbs are – unlike those accompanying ideophones – not necessarily synchronized.18 They are deictic, having the speaker as the origo or the reference object from where the motion departs or where it is directed to. Consequently, the gestures are typically pointing gestures. In addition, a motion event is almost always accompanied by gaze, which is directed to the outside source or goal of the movement.

3.1 Distribution of motion events in the sample texts

27The occurrence of motion events and the distribution of their components over the various modes of representation in the two longer sample texts, a mythological narrative and a task description, are summarized in Table 1.

28The first column of Table 1 provides a classification based on the conceptual components (manner, path, etc.) expressed in the main element of the construction which is given in parentheses. The second column lists the linguistic elements in which the conceptual categories are expressed (e.g., V, IDEO, PP) as well as the adjuncts or arguments (e.g., V (Mo&P) + PP (goal)). The third column codes the gesture that accompanies the construction, and in the fourth and the fifth column the distribution of occurrences in the two genres is given.

29The table is divided into five rows according to the conceptual categories encoded in the main linguistic element. One of these categories has two sub-categories due to the fact that ‘motion & manner’ can be either encoded primarily in a finite verb or in an ideophone.

  • 19 Since in Awetí discourse there is no perceptible difference between finite verbs and verbs that are (...)
  • 20 In Awetí discourse participants are usually encoded by verb prefixes and in some cases additionally (...)
  • 21 Strictly speaking, none of the components of a motion event are expressed linguistically by the dei (...)

30Further sub-categories are given in the column listing the overall linguistic expressions. Verbs19 encoding ‘motion & path’ (Mo&P) or ‘motion & manner’ (Mo&Ma) are either the only linguistic means to represent the motion event, or occur in combinations with postpositional phrases (PP) indicating ‘goal’ or ‘source’. Verbs of both types which occur without gestures are listed in additional sub-categories. There are further sub-categories for motion verbs in combination with full noun phrases (NP) referring to the moving ‘figure’.20 There is another sub-category for motion verbs combined with time adverbs. Five sub-categories are listed for ideophones. They can either precede or follow a clause with a verb referring to the same motion event (IDEO, V or V, IDEO). They can represent a motion event all by themselves (IDEO) or – as in one case – occur in combination with a PP referring to the ‘ground’ component (nãpeput ywã ‘on their former path’). There is further one instance of an ideophone following a locative non-verbal clause in which the PP (picicleta ‘apo ‘on the bicycle’) provides information on the ‘manner’ of motion. Other motion events are represented by the adverb nãtsu (‘like this’), deictically introducing ‘manner’,21 by specific hortative expressions (nawỹj / nã me ‘let’s go’) and by non-verbal clauses (NP, PP).

  • 22 The additional numbers in parentheses indicate that the corresponding gesture was not (clearly) vis (...)
  • 23 This category is problematic (see footnote 27).

Category

Linguistic expression

Gesture

Mythological narrative

Task description

1. Motion

V (Mo&P)

path

74

43 (+5)22

and Path

(verb)

V (Mo&P) +

   PP (goal)

path

16

9 (+1)

V (Mo&P) +

   PP (source)

path

1

---

V (Mo&P) +

   NP (figure)

figure

2

---

V (Mo&P) +

   ADV (time)

time

2

5

V (Mo&P)

---

1

5

2.1 Motion

V (Mo&Ma)

path

1

7

and Manner

(verb)

V (Mo&Ma) +

   PP (goal)

path

---

1

V (Mo&Ma) +

   NP (figure)

figure

---

1

V (Mo&Ma)

---

---

3

2.2 Motion

V, IDEO

path&manner

3

2

& Manner

IDEO

path&manner

1

1

(ideophone)

IDEO +

   PP(ground)

path&manner

1

---

V, IDEO

path

1

2

PP (manner), IDEO

path

---

1

3. Motion, Manner

ADV (‘like this’)

path

---

4

& Path23

ADV (‘like this’)

path&manner

---

1

4. Motion & Path

(hortative expression)

nawỹj / nã me

(‘let’s go!’)

path

3

2

5. Figure (NP)

NP (figure)

path

---

2

6. Goal (PP)

PP (goal)

path

4

---

Total

111

89 (+6)

Table 1. Distribution of motion components over elements

31The results of the quantitative analysis presented in Table 1 can be summarized as follows. In the two sample texts most motion events were accompanied by a gesture. Only four percent of all motion events – all of them consisting of simple motion verbs – are not gestured (1 token in the myth, 8 tokens in the task description).

32The table further indicates that motion events are most frequently expressed by ‘motion & path’ (Mo&P) verbs. Sixty percent of all 206 motion events in the sample texts (67% of the myth, 51% of the task description) consist of these verbs without any further constituent, 13% percent have an additional PP to encode a ‘goal’ or ‘source’ component of the motion event. Within these, the verbs to (‘go away’) and ut (‘come (back)’) predominate; see examples (4a) and (4b). Mo&P verbs are consistently accompanied by a ‘path’ gesture, independent of whether they are modified by a locative PP or not; see examples (4a), (4b), and (9b). In two cases iconic gestures depicting a ‘figure’ component co-occur with motion events of this type. This happens wherever a participant is newly introduced into the discourse, as will be illustrated by example (5) further down. There are further seven instances, where ‘time’ information is gestured instead of a motion component. In these contexts a pointing gesture to the position of the sun, as shown in example (9) below, is necessary to disambiguate a deictic adverb in the same clause.

33The linguistic as well as the gestural encoding of ‘manner’ is relatively rare in Awetí. Verbs providing information on ‘motion & manner’ (Mo&Ma), constituting only six percent (9 tokens) of all motion events, are also accompanied by a ‘path’ gesture, except for one case, where a newly introduced participant is gesturally depicted instead.

34Twelve motion events in the sample texts are represented by ideophones. This is not a high number in comparison to the motion events expressed by verbs, since ideophones are primarily used as expressive means to foreground decisive moments in narratives. Ideophones invariably provide a ‘manner’ component, which – with three exceptions, examples (7), (8), and (9c) – is also depicted gesturally. In the utterance where an ideophone is preceded by a non-verbal clause formed from a PP indicating ‘manner’ of movement the gesture accompanying the ideophone is not as elaborated as in examples where there is no ‘manner’ component otherwise expressed, i.e., the gesture here only conveys ‘path’ information; compare (9c) with (9d) below.

35Other strategies to refer to motion events in Awetí are used to a lesser degree but are invariably produced with a ‘path’ gesture directed away from the speaker. This applies to the five utterances of hortative expressions such as nawỹj (‘let’s go!’), occurring in both sample texts, by which interlocutors are asked to move together with the speaker (category 4).

36In the descriptive text the speaker further uses five times the deictic adverb nãtsu (‘like this’), which, similar to temporal adverbs, needs to be referentially disambiguated by a gesture. In this linguistic environment ‘manner’ of motion may – apart from the ‘path’ component – additionally be encoded in the gesture, as illustrated by example (9e) below.

37There are altogether six gestures in the two texts which, by providing the ‘motion’ and ‘path’ component on their own, replace a finite verb. In two of these examples of non-verbal clauses the ‘figure’ is expressed by a noun phrase (category 5), in four examples postpositional phrases indicate the goal (category 6).

38This short quantitative analysis of two texts does not claim to have statistical relevance. Nevertheless, the results obtained here regarding the gesturing of motion components in combination with specific linguistically given components can be regarded as tendencies which manifest themselves in a similar way in other texts in the Awetí corpus.

3.2 Discussion of examples

  • 24 Only example (11) is from an autobiographical narrative told by a different speaker.

39In this subsection the distribution of motion components over finite verbs, ideophones and gesture will be illustrated by examples from the two sample texts,24 and observable regularities will be discussed. Although ideophones are not as frequent as verbs and gestures in the context of motion events, they have been included in this discussion for several reasons. On the one hand, they represent a distinct mode of representation and are closely linked to co-speech gesture and thus provide further information on the interplay of the different modes. On the other hand, ideophones depict ‘manner’ which – after ‘motion’ and ‘path’ – is a central conceptual component of an Awetí motion event.

3.2.1 Motion components conveyed by gestures and finite verbs

40An example of a motion event expressed by the verb to (‘go’) is given in (4a):

(4)

a.

O-

to

ti

a’yn.

3-

go.vi

EVID

PART

rh

~~~~

***********

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-

‘He went away.’ (tal_kakaja, 0102.608)

41The gesture is performed rapidly, whereby the stroke coincides with to, the second syllable of the finite verb, which also receives primary stress in the utterance. The gesture stroke is captured in Figure 1.

Figure 1. Gesture ‘go’ with verb in example (4a)

42In (4b) from the same sample text the motion verb to is modified by the postpositional phrase ko kyty (‘to his field’), indicating the goal. As shown by the gesture stroke in Figure 2, an (almost) identical gesture, away from the speaker, is produced.

(4)

b.

O-

to

ti

ko

kyty

a’yn.

3-

go.vi

EVID

field

to

PART

rh

~~~~

*****

-.-.-.-.-.-.-.-.-.-.-.-

‘He went (away) to his field.’ (tal_kakaja, 1061.355)

Figure 2. Gesture ‘go to the field’ in example (4b)

  • 25 This has not yet been systematically checked in Awetí discourse. However, only part of the motion g (...)

43The same kind of gesture away from the speaker further occurs in utterances with the hortative expression nawỹj (‘let’s go’). The fact that these gestures away from the speaker are not always performed with the same hand and into the same direction suggests that Awetí speakers have an absolute orientation in space, indicating different geographical locations.25 The difference in linguistic information between example (4a) and (4b), with and without a PP, further shows that the pointing gesture by itself can have the communicative function of indicating a goal.

44The following example (5) and the gesture stroke in Figure 3 illustrate the gestural depiction of a ‘figure’ component of the motion event. The gesture in this case is time-aligned with the NP.

(5)

W=

e-

pokytsap

n=

ezu(t)

-tu

ti

(a)’yn.

3coref=

REL-

saw

3=

bring

-NOM

EVID

3PRO

PART

bh

~~~~~~~~~~~~~~~~~~~~~*****-.-.-.-

‘He brought his (own) saw.’ (tal_kakaja, 0098.611)

Figure 3. Depiction of ‘figure’ component in example (5)

45Full NPs in Awetí are mostly used in narrative discourse to (re)introduce a participant. Given participants tend to be unexpressed. In the sample texts the gesturing of the ‘figure’ component happens wherever the referent of an NP is new information.

3.2.2 Motion components conveyed by gestures, finite verbs and ideophones

  • 26 For discussion and examples see Reiter (2013), chapter 4.
  • 27 Compare the path gestures in (7), (8) and (9c) with the path&manner gestures in (6), (9a) and (9d) (...)

46Ideophones in Awetí frequently co-occur with gestures with which they are fully time-aligned, so that one may hypothesize that the two analogous modes of representation are conceptually linked. However, throughout the available Awetí video data it could be observed that ideophones in motion events do not co-occur with the type of iconic gestures that usually accompany ideophones when depicting other activities.26 Instead, they occur with the same pointing gestures that are used with motion verbs.  The prominent semantic contribution of these gestures is the ‘path’ component and thus considerably different from the meaning component encoded by ideophones, which is ‘manner’ of motion. Only where ‘manner’ information seems to be important in discourse, it tends to be additionally expressed in the gesture.27

47The information conveyed by pointing gestures co-occurring with ideophones depicting motion events correspond to the ‘path’ information that may also be encoded in the meaning of an adjacent verb (e.g. to – away from the speaker, ut – towards the speaker). This is exemplified by (6) with the verb to (‘to go (away)’) and the ideophone tirik (‘person walking on dry leaves (in the forest)’).

(6)

Nã=

to

-tu

me,

tirik tirik tirik tirik tirik

tyk

-yhy,

3=

go.vi

-NOM

PART

IDEO.walk.through.thicket

IDEO.stop

-INT

rh

~~~~~~~~~~~~~~~~~~~~~~~~~

*~*~*~*~*~*~*~*~*~

********

-.-.-.-

kaawatu

tup

-(p)u

ti

nã.

forest

see.vt

-NOM

EVID

3PRO

‘He walked through the thicket, stopped (suddenly), (when) he saw a (good) forest.’

(tal_kakaja, 0108.446)

  • 28 Action nominals of this type (somewhat simplifying) can be regarded as functional equivalents to fi (...)

48In (6) the preparation phase of the gesture sets in with the utterance of the action nominal nãtotu (‘he went’),28 and a clear stroke coincides with the first syllable of tykyhy, as captured in Figure 5. The recovery phase is concluded with the utterance of the remaining syllables of this last ideophone. During what is marked as a phase, where it is difficult to decide whether it is still part of the preparation of a gesture or its stroke, the arm moves slowly upwards, as shown in Figure 4, rhythmically accompanied by several repetitions of tirik. Syntactically, the ideophones occur after the clause-final particle me, so that they form two independent clauses (for the two distinct events walking and stopping). The gesture unit, however, extends over the clause boundary of the utterance. That the gesture expresses two semantic components, ‘path’ as well as ‘manner’, becomes obvious when comparing it to the rapid performance of the pointing gesture at the utterance of the finite verb oto (‘he went’) in (4a) and (4b) above.

Figure 4. Gesture ‘go’ with ideophone in example (6)

Figure 5. Gesture stroke at tykyhy (‘stop’) in example (6)

49In example (7) the motion verb ut (‘come’) is used together with a corresponding pointing gesture of a specific type, shown in Figure 6, which is repeated at the utterance of the ideophone tyk tyk (‘(person) walking (on the path)’) in the immediately following clause, as captured in Figure 7.

(7)

O-

ut

ti

n=

atitoza

a’yn,

tyk tyk.

3-

come.vi

EVID

3=

mother-in-law

PART

IDEO.walk

bh

************

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-

~~~*****

‘His mother-in-law came, (walking) tyk tyk.’ (tal_kakaja, 0576.926)

Figure 6. Gesture ‘come’ with verb in example (7)

Figure 7. Gesture ‘come’ with ideophone in example (7)

50The speaker lifts both hands rapidly from his lap before he starts talking and reaches the stroke at uttering the verb. The recovery phase ends after the first syllable of natitoza (‘his mother in law’), when the hands are again positioned in the speaker’s lap. As indicated in the gesture line in example (7) the gesture is repeated, when the ideophone is produced. After that, without a recovery phase, a different gesture follows to depict another event (not represented in the example). The gesture in (7), contrasting with the one in (6), is produced in a very rapid manner, and the fact that it is repeated, when the ideophone is uttered, permits the hypothesis that the same co-speech gesture is generally used for both verb and ideophone related to the same motion event. Unlike the interaction between the ideophone tirik (‘person walking in the forest’) and the gesture in example (6), there is no correspondence in rhythm or rate between this gesture and the ideophone tyk tyk (‘person walking (on the path)’). One reason may be that the ideophone used in (7) is less marked with respect to ‘manner’ than the one in (6), since tyk tyk is a kind of ‘default’ ideophone for human locomotion and has the additional ‘manner’ component ‘on the path (on clear ground)’ only when contrasted with another manner of motion; see example (9a). It must further be added that the verb ut (‘come’) normally co-occurs with a gesture which is only performed with one arm. The two-sidedness of the gesture here may indicate some emphasis, since in the narrative the appearance of the mother-in-law marks a turning point.

  • 29 Another possibility is that paddling in the context of fishing is seen as the default manner of loc (...)

51Another hypothesis for not depicting ‘manner’ in the gesture, for which there is additional evidence in example (8), may be that it is not a central piece of information in the respective discourse context. This example suggests that in gestures co-occurring with ideophones depicting motion events the ‘manner’ component is only expressed when this information plays a role in the development of the narrative. The speaker describes how men, when they go fishing, set out in their canoes to reach the place where the fish are caught. But although one would expect the paddling of a canoe, depicted in the ideophone, to be marked with regard to manner of locomotion, it is not depicted by a manual gesture in any way. This may be due to the fact that the utterance is immediately followed by a longer sequence in which the speaker focuses on the topic of fishing, describing and gesturally depicting the killing of different types of fish with various methods.29

(8)

Ozo=

to

-tu

ozo=

tetang

-kaw

a’yn.

pywpywpyw.

Pira’yt.

1pl.ex=

go.vi

-NOM

1pl.ex=

embark.vi

-GER

PART

IDEO.paddle

fish

lh

**-.-.-.-.-.-.-.-.-.-

~~~~~~~~~~~***-.-******-.-.-

‘We went to embark, (paddling) pywpywpyw. (There was) fish.’ (awu_fishing, 0378.83)

52When the speaker utters the action nominal ozototu (‘we went’), the pointing gesture, performed with his left arm, has already reached its stroke position, as indicated in Figure 8, and the hand is immediately withdrawn to a rest position on the speaker’s knee. This movement is concluded with the utterance of the third syllable of the action nominal.

Figure 8. Gesture stroke ’go’ with verb in example (8)

53The second pointing gesture sets in at the clause final particle a’yn. The stroke, captured in Figure 9, is time-aligned with the last syllable of the reduplicated ideophone pywpywpyw (‘paddling’). The arm is slightly withdrawn before the next word, and the utterance of pira’yt (‘fish’) is accompanied by another beat or stroke, after which the hand is lowered to its rest position.

Figure 9. Co-speech gesture with pywpyw in example (8)

54Both gestures, the one co-occurring with the action nominal and the one with the ideophone, are identical, depicting the ‘path’ (and possibly a ‘goal’) but not the (irrelevant) ‘manner’ component of the motion event.

3.3.3 Motion and time information depicted by gestures and ideophones

  • 30 Square brackets indicate that the respective (part of a) word was not clearly audible.

55In an earlier sequence from the same descriptive text about fishing practices, given in (9), pointing gestures accompanying motion events are combined with pointing gestures towards the position of the sun, where they have the function of specifying the deictic adverb (‘now’). In addition, manner of motion, slow walking versus fast bicycle-riding expressed by ideophones, plays a role in this sequence and is consequently depicted by gesture.30

(9)

a.

ozo=

[tem

-pu],

tyk tyk tyk tyk tyk tyk tyk.

now

1pl.ex=

go.out.vi

-NOM

IDEO.walk

lh

~1*******************************2****************

‘At this time (early) we went out, (walking) tyk tyk tyk tyk tyk tyk tyk.’

b.

ozo=

totem

-pu

tam

ywo

me,

katu

me.

now

1pl.ex=

arrive.vi

-NOM

village

at

PART

now

INT

PART

lh

3*******************************4*****************************-.-.-.-.-

‘At this time we arrived in the village, at this very moment.’

c.

itã

jatã

pisicleta

‘apo,

tyry.

now

TOP

DEM

bike

on

IDEO.ride

lh

(not visible)

~~~~~~~~5**-.-.-.-~~~~~~~~~~6*****~~~~~

‘Now it (this) is (going) on bicycle, tyry(ry).

d.

[we]ne

totem

[-pu]

tyryry

[me].

now

still

arrive.vi

-NOM

IDEO.ride

PART

lh

7**************-.-.-.-.~~~~~~~~8*********************

‘Now one still arrives (at midday), tyryry.’

e.

Nã=

tsu

-e’ym

mote

wian

ne.

3=

like

-NEG

long.ago

still

PART

lh

9*************************-.-.-.-.-.-.-.-.-.-.

‘It wasn’t like that in the past.’ (awu_fishing, 0353.113)

56The sequence contains three instances of two ideophones, one to depict the motion event of walking (tyk tyk tyk), the other two the bike-riding (tyry, tyryry). In addition, there are three instances of action nominals, which refer to motion events (tempu ‘go out’, totempu ‘arrive’). In two cases these have a relation to the directly following ideophones in separate clauses, which add a ‘manner’ component similar to adverbs in English (in utterance (9a) and (9d)). In one case the action nominal is the predicate of a clause without an ideophone in the immediate syntactic environment (in utterance (9b)).

57Altogether nine visible gestures, as marked in the glossing line, occur in the sequence. In (9a) the speaker, with his left hand, points up at (‘now’) to indicate by the angle of his arm the position of the sun at the time when people were leaving the village to go fishing. The pointing is reinforced by a gaze in the same direction, as captured in Figure 10. The same hand then describes a slow, arc-shaped movement towards a point in front of the speaker, depicting the movement of the sun together with the slow movement of the people who take a long time walking. Although it departs directly from the pointing gesture indicating time, this movement is analysed as a second gesture – completely synchronous with the ideophone – which combines ‘manner’ and ‘path’ of a motion event. The ideophone tyk tyk is formally the same as in example (6), depicting human beings walking on a path, but it is marked in discourse by being repeated several times, indicating that the walking is extended in space or time, and by being contrasted with tyryry, an ideophone for bike-riding, two clauses later.

Figure 10. Gesture 1 (sun position) in example (9a)

58When uttering the second (‘now’) in (9b), the speaker accelerates his movement for a short moment to perform another pointing gesture, shown in Figure 11, and then holds the hand at this position for the remainder of the utterance. This third gesture again makes a different semantic contribution to the utterance than the second gesture in pointing to the (low) position of the sun at the time when people traditionally came home from fishing. The gesture here clearly provides spatial information which is not contained in the deictic adverb (‘here’, ‘now’), while the hold of the gesture expresses a ‘path’ component of the motion event referred to by the verb totem (‘arrive’). At the utterance of the postpositional phrase tam ywo (‘at the village’), indicating the goal, the speaker additionally points with three fingers of the same hand to the front, while his index finger continues to point at the sun position. This is shown in Figure 12.

Figure 11. Gesture 3 (sun position) in example (9b)

Figure 12. Gesture 4 (pointing with three fingers) in example (9b)

59In utterance (9c) there are two visible gestures. The first, an arc-shaped movement to the left, is prepared from the second syllable of jatã (‘this’), reaching its stroke at the first two syllables of pisicleta (‘bicycle’). Here the gesture provides a ‘path’ component to a motion event which is not otherwise expressed in the locative non-verbal clause. The stroke is captured in Figure 13. Another stroke occurs during the utterance of the ideophone tyry. The speaker points upwards, holding his left arm in a right angle. By drawing this arm to the centre above his head he develops the seventh gesture in utterance (9d), which indicates the highest point of the sun at midday. This is shown in Figure 14.

Figure 13. Stroke of gesture 5 in example (9c)

Figure 14. Stroke of gesture 7 in example (9d)

60With the utterance of the next ideophone tyryry (‘bike-riding’) the speaker produces another gesture, drawing a full circle in the air. This gesture depicts by its circular movement the wheels or movement of a bicycle, i.e., the ‘motion’ and the ‘manner’ component. This gesture is illustrated in Figure 15.

Figure 15. Gesture 8 (bike-riding) in example (9d)

  • 31 I decided to analyse this sequence at the transition from (9d) to (9e) as containing two strokes of (...)

61When uttering the adverb nãtsu(e’ym) (‘(not) like this’) in (9e) the speaker describes another arc-shaped movement to the left. Although it departs directly from the circle-shape it is analysed as a ninth gesture, since it is not the depiction of a shape any more but the indication of a smooth outward movement, whereby the hand gradually opens, palm turned to the observer.31 After the stroke, extending to the first syllable of mote (‘in the past’), the hand is completely opened and a turn of gaze follows as can be seen in Figure 16. This gesture, again, is different in function since it provides ‘manner’ and ‘path’ information, which is not given by the deictic adverb nãtsu.

Figure 16. Turn of gaze after stroke of gesture 9 in example (9e)

62The sequence indicates that the encoding of a ‘manner’ component together with (or even without) the usual ‘path’ component in a gesture occurs, when this information is central in discourse. In this case the speaker compares the slow walking of earlier times (gesture 2) to the riding of bicycles nowadays (gesture 8 and 9), by which the duration of the fishing trip can be considerably shortened. This corresponds to the slow motion of the gesture accompanying the repeated ideophone tirik in example (6) where walking for a long time through the thicket until reaching an open space also plays an important role in the narrative. In (9d), bicycle-riding as a modern technique is further put into focus by an iconic gesture depicting the shape of a wheel which co-occurs with the ideophone tyryry (‘bike-riding’).

63There are further three occurrences of the deictic adverb immediately preceding a motion verb in the same utterance. In all three cases, in (9a), (9b), and (9d), the gesture indicating the sun’s position is either held (as in (9b)) or further developed into a ‘path’ gesture. During the hold of the gesture in (9b), the middle, ring and little finger point into the direction of a different place in front of the speaker at tam ywo (‘at the village’), while the index finger continues to point at the sun. In other occurrences in the corpus where the two gestures may enter in conflict, e.g., if pointing towards the sun is followed by a verb normally accompanied by a gesture directed towards the speaker, the time information but not the motion event is gestured. The reason must be seen in the specific usage of gestures with deictic expressions as objects of reference in otherwise incomplete sentences (see Kendon 2004: 175).

64 Another deictic expression, nãtsu (‘like this’) in (9e) is accompanied by a ‘path’ and ‘manner’ gesture. In this case the motion event is entirely expressed by this combination of adverb and gesture. Gesture five in (9c) additionally illustrates that a gesture, here in combination with a postpositional phrase referring to a ‘manner’ component, may express by itself the ‘motion’ and ‘path’ components of a motion event, even though this clause is immediately followed by an ideophone referring to manner of motion.

65The sequence shows that the interplay of language, ideophones and gesture is influenced by the discourse prominence of the respective information (such as ‘time’ and ‘manner’). Gestures (and arguably also the ideophone in (9c) and (9d) are further used to add information to deictic temporal and manner adverbs.

3.3.4 Additional referential meaning conveyed by gestures and ideophones

66The following example (10) illustrates that both gestures and ideophones can add further meaning components to an utterance. A gestural depiction of the ‘manner’ component in discourse-prominent ideophones may include the dimension of terrain. In this case the pointing direction of the gesture co-occurring with the ideophones is upwards, so that the depicted aspect of the manner of movement here is the information “in the air”. Example (11), in contrast, illustrates that gestures accompanying the same motion verbs without ideophones do not depict any ‘manner’ component.

67The utterance in (10) contains the motion verb atap (‘cross sth.’) and an ideophone in the immediately following clause, which depicts the ‘undulating’ flight of a woodpecker, characterized by a series of rapid wing beats in alternation with a bound.

(10)

Wej-

t-

atap

ti

me,

jatã

takururu,

3-

EPEN-

cross.vt

EVID

PART

DEM

cry.of.woodpecker

rh

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

popopopo popopopo popopopo popopopo popopopo popopopo.

IDEO.flying.of.woodpecker

**************************************************

‘It crossed (the field), this (one that makes) takururu, (flying) popopopo popopopo.’ (tal_kakaja, 0456.181)

68The repetition of an unusual number of ideophones, in addition, indicates duration, signaling that the bird in the narrative flies for a long time (and still does not reach the end of an enormous field planted with manioc).The specific manner of flying is additionally expressed by the prosodic shape of the utterance of the ideophone, as illustrated by the sound chart in Figure 17. The ideophone is rhythmically structured into six units of four syllables which – from the second to the sixth – decrease in loudness, while a second decrescendo is formed within each of these units. This narrative technique is used to indicate increasing distance. Each of the six units, depicting the beating of the bird’s wings, is followed by a short pause, by which the lack of motion during the bound is indicated.

Figure 17. Waveform and pitch of example (10)

69Figure 18 shows the end position of the preparation phase. The left arm of the speaker is fully flexed. He holds it at shoulder level, his elbow and gaze indicating the direction of the movement to the left. This shape of the preparation phase is held during the whole clause that precedes the ideophone.

Figure 18. End of preparation phase in example (10)

70Only when the speaker starts to utter the ideophone with its multiple repetitions the stroke of the gesture occurs. It consists of an arc-shaped movement with the left hand, index finger pointing, from a position near the right shoulder to the left side of the speaker. The highest point of the movement is captured in Figure 19, the moment, when a following clause sets in, in Figure 20.

Figure 19. Gesture ‘fly’ in example (10)

Figure 20. End of gesture in example (10)

  • 32 This suggests that the gesture depicting the event of ‘crossing’ may be conventionalized in Awetí.

71In order to confirm that the gesture co-occurring with the verb atap (‘cross sth.’) includes both components, i.e., a flexed elbow in the preparatory phase as well as pointing with the index finger,32 and that the slow arc-shaped movement semantically corresponds to the ideophone, an example from an autobiographical narrative with the same verb but without an ideophone is given in (11).

(11)

I=

topa(t)

-zoko

-tu,

i=

pe

-put

a-

t-

atap

mejka

‘e.

1sg=

be.lost.vi

-IMPF

-NOM

1sg-

path

-NPST

1sg-

EPEN-

cross.vt

EPIST

PART

rh

~~~~~~~~~~~

******************************

‘I was getting lost. I crossed my former path without perceiving it.’ (kup_autobiog, 1159.328)

72The gesture which co-occurs with the utterance containing the finite verb atatap (‘I crossed s.th.’) starts as shown in Figure 21. Similar to the preparation phase of the gesture captured in Figure 18 the elbow of the left arm points to the left side, while the outstretched hand is positioned at the right side. However, unlike in (10) the gesture stroke sets in immediately, as illustrated by Figure 22, in that the left hand moves rapidly in a horizontal line in front of the body to the left side, ending with a pointing of the index finger into that direction. With the following clause the speaker withdraws his hand to a rest position.

Figure 21. End of preparation phase in example (11)

Figure 22. Gesture stroke ‘cross’ in example (11)

  • 33 Shintel, Nusbaum & Okrent (2006) present experimental evidence for the hypothesis that speakers nat (...)

73The last two examples illustrate again that pointing gestures when accompanying ideophones may differ from the pointing gestures co-occuring with motion verbs in conveying ‘manner’ information, here with regard to the terrain the motion takes place in. In (10) the ideophone is prominent in discourse, as is further conveyed by multiple repetition and prosodic markedness. In addition, both ideophones and gesture as analogous modes of representation can be interpreted here as non-arbitrarily mapping form and meaning. The ideophone, by its rhythmic structure,variation in fundamental frequency, and repetition, acoustically depicts the long flight of a specific kind of bird, while the co-occurring arc-shaped gesture above the speaker’s head provides information on the terrain the motion takes place in.33

  • 34 See Reiter (2013) for a further discussion of sound-symbolism in ideophones.

74Ideophones, in addition, often have sound-symbolic qualities. An impression of this may be captured by comparing the ideophones in examples (6), (7), (8), and (10). While the stepping on firm ground by humans is expressed in a plosive /k/ in the coda of the last syllable in tirik (‘walking on dry leaves’) or of the only syllable in tyk (‘walking on a clear path’), the ideophones depicting motion in the air or in water have a vowel, as in popopo, or semi-vowel, as in pywpywpyw, in the coda. Differences in terrain influencing the manner of motion can further be depicted by using either one syllable, as in tyk, or two, as in tirik.34 Gestures and ideophones can thus both depict further meaning components of a motion event which are not conveyed by linguistic means.

4. Conclusions

  • 35 See McNeill & Duncan (2000: 150) who observed a similar pattern in English, but a completely differ (...)

75Our analysis of two fully-coded sample texts suggests that nearly all motion events in Awetí are gestured. Moreover, the gestured component is almost exclusively ‘path’, independent of whether this information is also conveyed verbally or not. These gestures usually co-occur with the verb and, if present, with an ideophone. In contrast to Mandarin and similar to Spanish, ‘path’ verbs are very frequent in Awetí, while ‘manner’ verbs or ideophones depicting this component are relatively rare. This provides further evidence for Chui’s (2009: 1775) hypothesis that the ‘path’ component – independent of language structure – is prominent in the mental conceptualization of motion events, a fact which was initially predicted by Talmy (2007: 139) who defined ‘path’ as a ‘core component’ of motion events. With regard to information state Chui’s hypothesis that ‘figure’ and not ‘path’ information is gestured, if the participant is newly introduced in the discourse, could be corroborated by Awetí data. For ‘manner’ it could be shown that this component is gestured additionally to ‘path’, only if expressed in a co-occurring ideophone. In addition, this information needs to be prominent in discourse, e.g., in terms of contrast between one type of motion and another or in order to convey additional spatial and/or temporal information.35 Another type of motion event which needs to be complemented by a gesture and accordingly may also contain a gestured ‘manner’ component is formed by the deictic adverb nãtsu (‘like this’). Similarly, ‘time’ information is conveyed by a deictic adverb (e.g., ‘here, now’) in combination with a pointing gesture to the respective position of the sun. Since ‘time’ seems to be the dominant concept, the pointing to the sun may either prevent the gesturing of motion or be combined with a ‘path’ gesture.

76The close interplay between gestures and ideophones further raises the question whether these may be actually two components of one analogue mode of representation, corroborating the claim that ideophones are in fact vocal gestures rather than an element of spoken language. As could be illustrated by the examples, gestures and ideophones tend to be fully synchronized. The meaning components they convey are mostly complementary and not – as in the relation between gesture and ‘path’ verb – a parallel expression of the same component in a different mode, i.e., the gesture primarily depicts the ‘path’, while the ideophone conveys information on ‘manner’ of motion in a broad variety of ways (by repetition, prosodic means and sound symbolism).

77Especially in their combination with ideophones and linguistic elements other than verbs it could be shown that the embodiment of motion components by co-speech gestures in Awetí is not merely an externalization of cognition (as McNeill & Duncan (2000: 142) say), but also that the gestures are voluntarily enacted by the speakers to convey meaning. Awetí co-speech gestures clearly serve a communicative function, not only with otherwise incomplete deictic adverbs, but also with NPs and PPs in non-verbal clauses, where they add the ‘motion’ component, or with ideophones which do not encode ‘path’ by themselves. With regard to ideophones referring to activities other than motion, there is evidence that co-occurring gestures are usually produced with a high level of speaker awareness and are partially conventionalized (see Reiter 2013). The last two examples (10) and (11) also suggest such a conventionalization for gestures which are produced together with specific Awetí motion verbs. Most ‘motion’ gestures, however, do not show a specific recurrent form, i.e., the movement ‘away from origo’ or ‘towards origo’ can be performed in various ways, possibly influenced by the speaker’s absolute orientation in space, as discussed in some of the examples. Finally, in addition to providing information not given linguistically, gestures in Awetí can also visually foreground central meaning components in discourse, as in the case of the depiction of new participants or with regard to manner of motion. Further investigations into video data obtained from a variety of speakers in a controlled experimental setting will be necessary to sufficiently corroborate the preliminary results of this study.  

Haut de page

Bibliographie

Chui, K. 2009. Linguistic and imagistic representations of motion events. Journal of Pragmatics 41: 1767-1777.

Dingemanse, M. 2011. The Meaning and Use of Ideophones in Siwu. Doctoral Dissertation. Radboud Universiteit Nijmegen.

Franchetto, B. 2010. Bridging linguistic research and linguistic documentation: The Kuikuro experience (Brazil). In: Flores Farfán, J.A. & F.F. Ramallo (eds.) New Perspectives on Endangered Languages. Bridging gaps between sociolinguistics, documentation and language revitalization. Amsterdam/Philadelphia: John Benjamins, 49-64.

Gabas, N. Jr. 2006. Tupian Languages. In: Brown, K. (ed.). Encyclopedia of Language & Linguistics. 2nd edition. Vol. 13. Oxford: Elsevier, 146-150.

Güldemann, T. 2008. Quotative indexes in African languages: A synchronic and diachronic survey. Berlin: Mouton de Gruyter.

Güldemann, T. 2011. The history of quotative predicatives: Can lexical properties arise out of a grammatical construction? Paper presented at Linguistisches Kolloquium, 10/1/2011, Seminar für Afrikawissenschaften, Humboldt Universität zu Berlin.

Johnson, M. 1987. The Body in the Mind. Chicago: University of Chicago Press.

Kendon, A. 2004. Gesture: Visible Action as Utterance. Cambridge: CUP.

Kita, S. 1997. Two-dimensional semantic analysis of Japanese mimetics. Linguistics 35(2): 379-415.

Kunene, D.P. 1978. The Ideophone in Southern Sotho. Marburger Studien zur Afrika- und Asienkunde, Serie A, Band 11. Berlin: Dietrich Reimer.

McNeill, D. 1992. Hand and Mind. What Gestures reveal about thought. Chicago: University of Chicago Press.

McNeill, D. 2000 (ed.). Language and Gesture. Cambridge: CUP.

McNeill, D. & S.D. Duncan. 2000. Growth points in thinking-for-speaking. In: McNeill, D. (ed.) Language and Gesture. Cambridge: CUP, 141-161.

Nuckolls, J.B. 1996. Sounds like Life: Sound-Symbolic Grammar, Performance, and Cognition in Pastaza Quechua. New York/Oxford: OUP.

Nuckolls, J.B. 2001. Ideophones in Pastaza Quechua. In: Voeltz, F.K.E. & C. Kilian-Hatz (eds.) Ideophones. Typological Studies in Language 44. Amsterdam/Philadelphia: John Benjamins, 271-285.

Özyürek, A., S. Kita, S. Allen, R. Furman & A. Brown. 2005. How does linguistic framing of events influence co-speech gestures? Insights from crosslinguistic variations and similarities. In: Liebal, K., C. Müller & S. Pika (eds.) Gestural Communication in Nonhuman and Human Primates. Amsterdam/Philadelphia: John Benjamins, 199-218.

Reiter, S. 2010.Linguistic vitality in the Awetí indigenous community: A case study from the Upper Xingu multilingual area. In: J.A. Flores Farfán & F.F. Ramallo (eds.)New Perspectives on Endangered Languages. Bridging gaps between sociolinguistics, documentation and language revitalization. Amsterdam/Philadelphia: John Benjamins, 119-146.

Reiter, S. (2013). Ideophones in Awetí. Doctoral dissertation, University of Kiel.

Samarin, W.J. 1971. Survey of Bantu Ideophones. African Language Studies 7: 130-168.

Shintel, H., H.C. Nusbaum & A. Okrent. 2006. Analog acoustic expression in speech communication. Journal of Memory and Language 55: 167-177.

Talmy, L. 2007. Lexical typologies. In: Shopen, T. (ed.) Language Typology and Syntactic Description. Vol. 3. 2nd edition. Cambridge: CUP, 66-168.

Haut de page

Annexe

Abbreviations

1pl.ex – 1st plural exclusive; ADV – adverb; DEM – demonstrative; EPEN – epenthesis; EPIST- epistemic marker; EVID – evidential; GER – gerund; IDEO – ideophone; IMPF – imperfective; INT – intensification; Mo&Ma – motion and manner; Mo&P – motion & path; NOM – nominalizer; NP – noun phrase; NPST – nominal past; PART – clause-final particle; PP – postpositional phrase; PRO – pronoun; TOP – topicalizer; V – verb; vi – intransitive verb; vt – transitive verb

Haut de page

Document annexe

Haut de page

Notes

1 I have excluded the ‘continuation of a stationary location’ which Talmy (2007: 70) together with motion subsumes under a ‘Motion event’ since it is not relevant for the analysis presented here. Therefore, the semantic components, which only refer to the parts of an event actually involving motion, are not written with a capital letter as in Talmy’s terminology.

2 According to whether the ‘path’ component is encoded in or outside the main verb Talmy (2007) classifies languages as “satellite-framed” (e.g. English) or “verb-framed” (e.g. Spanish). It must be added that these classifications capture the predominant pattern, since languages usually have various possibilities to encode motion events which are used to differing degree.

3 21.6% of all motion events contained ground information which was gesturally represented in 26.4% of these instances. Figure information was given in 27.3% of the motion events and represented in 22.4% of these (see Chui 2009, Table 6).

4 A prerequisite for this hypothesis is that she assumes gesture to be closer to the conceptualization of an event than language structure. For Chui (2007: 1775) evidence is provided by the fact that gestures produced with typologically different languages depict identical components of a motion event.

5 I have avoided to classify the gestures considered in this study according to the McNeill (1992) classification as “representational“ (iconic), in opposition to “deictic” (pointing), “emblems“, and “beats“, because many of them do not neatly fit into one of the categories. Instead, they are classified according to the meaning components they convey. In terms of McNeill’s classification gestures depicting the ‘path’ and the ‘manner’ component of a motion event would represent a combination of the “iconic“ and the “pointing“ type.

6 See also http://www.unesco.org/new/en/culture/themes/endangered-languages/language-vitality/

7 See www.mpi.nl/DOBES/projects/aweti.

8 The documentation setting could be optimized, if members of the community themselves recorded the data in the absence of outside researchers (see, e.g., Franchetto 2010).

9 Other motion events were entirely expressed by gesture, either in combination with an NP as the ‘figure’ component or with a deictic adverb (‘like this’). Such occurrences, also listed in Table 1, are not in the focus of this paper and will not be illustrated by corpus examples.

10 See, e.g., Nuckolls (1996).

11 In each example the information in brackets after the free translation refers to the source file in the Awetí data corpus.

12 Thesegestures have iconic and deictic features to differing degree. As will be shown in the examples, gestures that co-occur with motion events are often a combination of “iconic” and “deictic”.

13 Since the overall number of ideophones in the video data used for this study is relatively low, this must be treated as a preliminary statement. Further analyses, e.g., of elicited narratives as used by Kita (1997), are needed in order to draw any valid conclusions.

14 Note that Kita (1997) only uses the term ”iconic gestures”. However, his descriptions of the gestures that co-occur with ideophones suggest that these ”iconic gestures” also have a deictic component (see footnote 9). See also Dingemanse (2011) who draws a distinction between “depictive” and “deictic” gestures co-occurring with ideophones depicting different events. According to his analyses depictive gestures are far more frequent than other gesture types in these contexts, but only little more than a third of the 174 ideophones in his corpus of Siwu (West Africa) are gestured at all.

15 See example (10) below.

16 These are the two most widespread types of motion verbs according to the typology established by Talmy (2007: 99, Table 2.2).

17 Different from their English translation equivalents go and come which are classified as deictic verbs not encoding a ‘path’ component, to and ut in Awetí imply that the location from which the movement departs or where it is directed to is, respectively, the home village of the speaker or of a protagonist whose perspective is taken in a narrative. By thus being invariable throughout the narrative with regard to this parameter, the two verbs are here classified as ‘path’ and not as ‘deictic’ verbs.

18 Further Awetí video data needs to be fully annotated in order to accumulate a significant number of ideophones in motion events and comply with the suggestion given by an anonymous reviewer of this paper to compare the correlation between gestures and ideophones with the one between gestures and motion verbs.

19 Since in Awetí discourse there is no perceptible difference between finite verbs and verbs that are formally action nominal constructions, both are treated here as ‘verbs’ (V).

20 In Awetí discourse participants are usually encoded by verb prefixes and in some cases additionally by personal pronouns. Full NPs mark discourse prominence of the participant referred to.

21 Strictly speaking, none of the components of a motion event are expressed linguistically by the deictic adverb. The conceptual components are expressed by non-linguistic modes, here by the gestures which form a unit with the deictic adverb, expressing either ‘motion’, ‘path’, ‘manner’ or all three components together. According to Güldemann (2011: 4) deictic adverbs of this kind (‘manner deictics’ in his terminology, see 2008: 318) are cross-linguistically a typical source for so-called ‘quotative indexes’, markers to introduce mimetic signs like gestures or ideophones into syntactic structure.

22 The additional numbers in parentheses indicate that the corresponding gesture was not (clearly) visible.

23 This category is problematic (see footnote 27).

24 Only example (11) is from an autobiographical narrative told by a different speaker.

25 This has not yet been systematically checked in Awetí discourse. However, only part of the motion gestures seem to be directed towards a geographical location.

26 For discussion and examples see Reiter (2013), chapter 4.

27 Compare the path gestures in (7), (8) and (9c) with the path&manner gestures in (6), (9a) and (9d) and (10).

28 Action nominals of this type (somewhat simplifying) can be regarded as functional equivalents to finite verbal predicates in Awetí (see footnote 24).

29 Another possibility is that paddling in the context of fishing is seen as the default manner of locomotion.

30 Square brackets indicate that the respective (part of a) word was not clearly audible.

31 I decided to analyse this sequence at the transition from (9d) to (9e) as containing two strokes of different gestures which immediately follow each other. Alternatively, a vocal (ideophone) and a co-occurring manual gesture are together introduced into discourse by the deictic adverb nãtsu(e’ym) (‘(not) like this’) (see footnote 27). The gesture would thus consist of both a circle shape (‘manner’) and a movement to the left (‘path’). The only problem to this alternative analysis is that the discourse particle me, following the ideophone, usually indicates the end of a clause.

32 This suggests that the gesture depicting the event of ‘crossing’ may be conventionalized in Awetí.

33 Shintel, Nusbaum & Okrent (2006) present experimental evidence for the hypothesis that speakers naturally and spontaneously use “analog acoustic expression” (here in the form of ideophones) to convey referential information, independent of whether this information is expressed verbally. The authors further show that the information conveyed exclusively through acoustic properties of speech is understood by listeners, thus confirming that they serve a communicative function.

34 See Reiter (2013) for a further discussion of sound-symbolism in ideophones.

35 See McNeill & Duncan (2000: 150) who observed a similar pattern in English, but a completely different one in Spanish, even though the latter is more similar to Awetí in its linguistic realization of motion events. In Mandarin, as stated by Chui (2009, Tables 3 and 5), the discourse-prominence of ‘manner’ does not influence its gesturing.

Haut de page

Table des illustrations

Légende Figure 1. Gesture ‘go’ with verb in example (4a)
URL http://cognitextes.revues.org/docannexe/image/765/img-1.jpg
Fichier image/jpeg, 20k
Légende Figure 2. Gesture ‘go to the field’ in example (4b)
URL http://cognitextes.revues.org/docannexe/image/765/img-2.jpg
Fichier image/jpeg, 20k
Légende Figure 3. Depiction of ‘figure’ component in example (5)
URL http://cognitextes.revues.org/docannexe/image/765/img-3.jpg
Fichier image/jpeg, 20k
Légende Figure 4. Gesture ‘go’ with ideophone in example (6)
URL http://cognitextes.revues.org/docannexe/image/765/img-4.jpg
Fichier image/jpeg, 20k
Légende Figure 5. Gesture stroke at tykyhy (‘stop’) in example (6)
URL http://cognitextes.revues.org/docannexe/image/765/img-5.jpg
Fichier image/jpeg, 20k
Légende Figure 6. Gesture ‘come’ with verb in example (7)
URL http://cognitextes.revues.org/docannexe/image/765/img-6.png
Fichier image/png, 208k
Légende Figure 7. Gesture ‘come’ with ideophone in example (7)
URL http://cognitextes.revues.org/docannexe/image/765/img-7.jpg
Fichier image/jpeg, 20k
Légende Figure 8. Gesture stroke ’go’ with verb in example (8)
URL http://cognitextes.revues.org/docannexe/image/765/img-8.png
Fichier image/png, 259k
Légende Figure 9. Co-speech gesture with pywpyw in example (8)
URL http://cognitextes.revues.org/docannexe/image/765/img-9.jpg
Fichier image/jpeg, 24k
Légende Figure 10. Gesture 1 (sun position) in example (9a)
URL http://cognitextes.revues.org/docannexe/image/765/img-10.jpg
Fichier image/jpeg, 20k
Légende Figure 11. Gesture 3 (sun position) in example (9b)
URL http://cognitextes.revues.org/docannexe/image/765/img-11.jpg
Fichier image/jpeg, 20k
Légende Figure 12. Gesture 4 (pointing with three fingers) in example (9b)
URL http://cognitextes.revues.org/docannexe/image/765/img-12.jpg
Fichier image/jpeg, 20k
Légende Figure 13. Stroke of gesture 5 in example (9c)
URL http://cognitextes.revues.org/docannexe/image/765/img-13.jpg
Fichier image/jpeg, 20k
Légende Figure 14. Stroke of gesture 7 in example (9d)
URL http://cognitextes.revues.org/docannexe/image/765/img-14.jpg
Fichier image/jpeg, 24k
Légende Figure 15. Gesture 8 (bike-riding) in example (9d)
URL http://cognitextes.revues.org/docannexe/image/765/img-15.png
Fichier image/png, 260k
Légende Figure 16. Turn of gaze after stroke of gesture 9 in example (9e)
URL http://cognitextes.revues.org/docannexe/image/765/img-16.png
Fichier image/png, 261k
Légende Figure 17. Waveform and pitch of example (10)
URL http://cognitextes.revues.org/docannexe/image/765/img-17.png
Fichier image/png, 88k
Légende Figure 18. End of preparation phase in example (10)
URL http://cognitextes.revues.org/docannexe/image/765/img-18.jpg
Fichier image/jpeg, 20k
Légende Figure 19. Gesture ‘fly’ in example (10)
URL http://cognitextes.revues.org/docannexe/image/765/img-19.png
Fichier image/png, 213k
Légende Figure 20. End of gesture in example (10)
URL http://cognitextes.revues.org/docannexe/image/765/img-20.png
Fichier image/png, 213k
Légende Figure 21. End of preparation phase in example (11)
URL http://cognitextes.revues.org/docannexe/image/765/img-21.jpg
Fichier image/jpeg, 24k
Légende Figure 22. Gesture stroke ‘cross’ in example (11)
URL http://cognitextes.revues.org/docannexe/image/765/img-22.png
Fichier image/png, 253k
Haut de page

Pour citer cet article

Référence électronique

Sabine Reiter, « The multi-modal representation of motion events in Awetí discourse », CogniTextes [En ligne], Volume 9 | 2013, mis en ligne le 21 mai 2015, consulté le 29 juin 2017. URL : http://cognitextes.revues.org/765 ; DOI : 10.4000/cognitextes.765

Haut de page

Auteur

Sabine Reiter

Federal University of Pará, Casa de Estudos Germânicos, Campus Universitário do Guamá, Setor Profissional / Pavião GP Altos, BR-66075-900 Belém, Brazil

Haut de page
  • Logo DOAJ – Directory of Open Access Journals
  • Logo ERIH PLUS | NSD
  • Logo AFLiCo – Association française de linguistique cognitive
  • Les cahiers de Revues.org