(In press) “Linguistic Typology in Motion Events: Path and Manner”. Anuario del Seminario de Filología Vasca ‘Julio de Urquijo’. International Journal of Basque Linguistics and Phylology. LINGUISTIC TYPOLOGY IN MOTION EVENTS: PATH AND MANNER Iraide Ibarretxe-Antuñano Pagazaurtundua, 3 - 2D. E-48980 Santurtzi. Bizkaia. Tlfnoa: 658713224 E-posta: [email protected] Linguistic typology in motion events: Path and Manner* IRAIDE IBARRETXE-ANTUÑANO University of Deusto – University of the Basque Country In each language only a part of the complete concept that we have in mind is expressed, and each language has a peculiar tendency to select this or that aspect of the mental image which is conveyed by the expression of the thought (Boas, 1911/1966: 38-39) Abstract Motion events are situations “containing movement or the maintenance of a stationary location” (Talmy 1985: 85). In the last twenty years there has been an increasing interest in the research of motion events. Much of this interest has arisen from the seminal work of Len Talmy and his lexicalisation patterns (1985, 1991, 2000), and that of Dan Slobin and his ‘Thinking for speaking’ hypothesis (1987, 1991, 1996a,b, 1997, 2000). In this paper we are going to analyse and compare motion events in three different languages: English, Spanish and Basque. Our goal is to compare these data under a critical eye so that we can (i) discuss the importance of the differences and similarities that exist in the expression of motion in these languages with respect to our language use, (ii) test the validity of Talmy’s typological dichotomy. 2 1. Introduction Motion events are situations “containing movement or the maintenance of a stationary location” (Talmy 1985: 85). Talmy’s (1985, 1991, 2000) typology divides languages into satellite-framed and verb-framed languages. The former expresses the semantic components of the motion event, Motion and Manner, conflated in the verb, and the Path in a satellite. The latter conflates Motion and Path in the verb, and expresses Manner in a separate expression. Slobin’s (1987, 1991, 1996a,b, 1997, 2000) ‘Thinking for speaking’ hypothesis applies this dichotomy to narrative and the issue of linguistic relativity. He argues that “lexicalisation patterns have consequences for the ways in which speakers focus on these components separately and in interaction, as reflected both in the lexical choice and the syntax of narrative discourse” (1997: 439). Taking Talmy’s and Slobin’s ideas as our theoretical basis, in this paper we are going to analyse and compare motion events in three different languages: English, Spanish and Basque. These languages are a good choice for comparison because they are different and similar at the same time. They are different because they belong to different language families –two Indo-European, Germanic (English) and Romance (Spanish), one non-Indo-European (Basque)-, but they are also similar because they share a Western cultural background, and in the case of Basque and Spanish a severalcentury-old contact. They are different because their lexicalisation patterns are not the same –English is satellite-framed, and Basque and Spanish verb-framed- but, as we shall see, there are also similarities among them, especially with respect to path and manner. 3 Our data come from previous analyses of motion events in the Frog stories (Berman and Slobin 1994). These are a collection of elicited narratives based on a wordless picture book, Frog where are you? (Mayer 1969), that tells the story of a boy who, together his dog, goes in search of his lost pet frog. The method for data gathering is very simple. The informant is told that s/he has to tell a story about a boy, a dog, and a frog. S/he is shown the book and allowed to look through the pictures. Then, s/he is asked to tell the story to the researcher who records or videotapes it. The informant follows the book picture by picture. English and Spanish data are mainly drawn from Slobin’s work, especially his 1996a paper ‘Two ways to travel’. Basque data are taken from Ibarretxe-Antuñano (2002, forthcoming) 1 . The main aim of this paper is not to show and discuss new data in reference to motion events –as we have just said, the data that we use are ‘borrowed’ from previous analyses. Our goal is to compare these data under a critical eye so that we can (i) discuss the importance of the differences and similarities that exist in the expression of motion in these languages with respect to our language use, (ii) test the validity of Talmy’s typological dichotomy. 2. Motion events and Language Typology Leonard Talmy describes a motion event as a situation “containing movement or maintenance of a stationary location” (1985: 61). According to this author, motion events are analysable into a set of six basic semantic elements or components, the first four constituting the central or ‘internal components’ while the last two are associated or ‘external co-event components’. These are: (i) ‘Figure’: the moving object; (ii) ‘Ground’: entity or entities that the Figure is moving in relation to; (iii) ‘Path’: the course followed (and trajectory) of the Figure; (iv) ‘Motion’: the presence of motion 4 per se; (v) ‘Manner’: the way in which motion is performed; and (vi) ‘Cause’: what originates the motion itself. Let us illustrate each of these components with some of Talmy’s classical examples (1985: 61): (1) The pencil rolled off the table Figure Motion Path Ground Manner (2) The pencil blew off the table Figure Motion Path Ground Cause In both cases the pencil plays the role of the Figure and the table that of the Ground which in these examples also expresses source of movement. The particle off functions as the Path. The verbs roll and blew express the Motion component. In addition, roll in (1) offers information about the Manner of motion, and blew in (2) about the Cause of motion. Talmy (1985, 1991, 2000) proposes that languages can be grouped together on the basis of how they encode the core information of a specific semantic domain onto syntactical and lexical structures. There are two distinct groups: those that allocate information in the verb and those that do so in some other elements called ‘satellites’ 2 . 5 In the case of motion events, from the six meaning components that play a role in the conceptual structure of a motion event, Talmy suggests that the basic information is precisely the motion of an entity along a path in a specified direction. In languages like German, the verb does not encode this information. Instead, it usually expresses the act of motion itself conflated with information about manner, that is, information about the way in which motion is performed. For example, verbs like rennen ‘run’, springen ‘jump’, krabbeln ‘crawl’. Path information is usually expressed in satellites such as (r)aus ‘out’, (r)ein ‘into’, (r)unter ‘down’, as in rennen raus ‘run out’. Languages like French follow the opposite strategy. Here, the core information is not expressed in a separate element but usually conflated with the verb, e.g. descendre ‘go down’, entrer ‘go in’. The encoding of manner is an optional choice in French, and thus, it is expressed in a separate element, e.g. entrer en courant ‘go in running’. Languages like German are called ‘satellite-framed languages’, and languages like French are called ‘verb-framed languages’. According to Talmy, languages can be classified into these two lexicalisation types on the basis of their ‘characteristic expression of motion’. By ‘characteristic’, Talmy (1985: 62; 2000b: 27) means that “(1) it is colloquial in style, rather than literary, stilted, and so on; (2) it is frequent in occurrence in speech, rather than only occasional; (3) it is pervasive, rather than limited –that is, a wide range of semantic notions are expressed in this type”. In the case of the languages under investigation in this paper, previous analyses 3 on their ‘characteristic expression of motion’ would place English as satellite-framed, and Spanish and Basque as verb-framed, as illustrated in (3), (4), and 6 (5), respectively. The Path component is in bold and the Manner of motion is underlined. (3a) The dog tumbles out of the window [B-S-94] (3b) An owl flew out of the hole in the tree [S-00] (4a) Del agujero salió un buho [S-00] of.the hole exited an owl ‘An own went out of the hole’ (4b) El perro salió corriendo [S-91] the dog exited running ‘The dog ran out’ (5a) eta ontzitikan atera intzan [IA-02] and jar:ABL:LOC exit:PERF inten.aux ‘And [the frog] went out of the jar’ (5b) eta erlauntzatik erle guztiak irten ziren hegaka [IA-02] and beehive:ABS bee all:ABS:DET:PL exit:PERF aux flying ‘And all the bees flied out from the beehive’ As it is characteristic of satellite-framed languages, the core information of the motion event is expressed outside the verb, in a satellite, the particle out in (3), 7 whereas the verb conflates motion and manner as in the verbs fly and tumble. In the frog stories, Slobin (1996a) reports that there are 123 different types of combinations between verbs and satellites. The lexicalisation pattern of verb-framed languages, on the other hand, follows the opposite strategy. The core information of the motion event is lexicalised in the verb. Thus, the verbs in these sentences, salir, irten and atera ‘exit’, convey both motion and path (outwards). In cases where the motion event also includes the semantic component of Manner this is usually expressed outside the verb, in a separate expression, a gerund in Spanish –corriendo ‘running’- and an adverb in Basque –hegaka ‘flying’- as we can see in (4b) and (5b) respectively. The information that these manner expressions add to the whole motion event is quite varied. It can refer to the following: (i) Motor pattern, e.g. Sp. corriendo ‘running’, rodando ‘rolling’, Bq. arinka ‘running’, arrastaka ‘dragging’, taka-taka ‘small steps’. (ii) Rate or speed of motion, e.g. Sp. de repente ‘suddenly’, a toda prisa ‘fast’, Bq. bizkor ‘quickly’, ziztu bizian ‘fast’. (iii) Means of transport, e.g. Bq. oinez ‘on foot’. (iv) Protagonist’s inner state 4 , e.g. Sp. asustado ‘scared’, enfadado ‘angry’, Bq. txintxo-txintxo ‘well-behaved’, zain-zain ‘watchful’. 8 As we can see from these examples, not only the information conveyed in these expressions is diverse, but also the grammatical category of these ‘separate expressions’. Adverbs (bizkor ‘quickly’), adjectives (asustado ‘scared’), gerunds (corriendo ‘running’), prepositional phrases (a toda prisa ‘fast’), and even sound symbolic expressions 5 (taka taka ‘small steps’) are used in this slot. 3. Thinking for speaking: language typologies in our language use The typological differences across languages described in the previous section show that languages have different syntactical-semantic preferences when they want to talk about motion. The detailed description of these preferences in several languages is an interesting exercise because it contributes to what authors such as Myhill (1992: 1-2) have called ‘typological discourse analysis’. That it to say, “the cross-linguistic study of the factors affecting the choice of one construction or another in a given language, taking the surrounding discourse context into consideration as having a crucial effect on this choice”. However, these language-specific patterns are not only important for their typological discourse implications –how speakers of different languages narrate the same story-, but also for their cognitive implications. As Berman and Slobin (1994: 612) argue, these discourse differences “suggest that the native language directs one’s attention, while speaking, to particular ways of filtering and packaging information”. With this idea in mind, Slobin (1987, 1991, 1996a,b, 1997, 2000) has put forward what he calls the ‘thinking for speaking’ hypothesis, a modified 6 version of the classical Sapir-Whorf debate on linguistic relativity in the first half of the twentieth century (Sapir 1924, Whorf 1940) based on psycholinguistic and typological research. Slobin himself explains the core motivation for this hypothesis in the following way: 9 The expression of experience in linguistic terms constitutes thinking for speaking – a special form of thought that is mobilised for communication. … We encounter the contents of the mind in a special way when they are being accessed for use. That is, the activity of thinking takes on a particular quality when it is employed in the activity of speaking. In the evanescent time frame of constructing utterances in discourse one fits one’s thoughts into available linguistic frames. ‘Thinking for speaking’ involves picking those characteristics of objects and events that (a) fit some conceptualisation of the event, and (b) are readily encodable in the language (Slobin 1991: 12) In other words, experience cannot be verbalised without having taken a specific perspective influenced, if not determined, by the typological characteristics and lexicalisation pattern of a given language. What we experience/perceive might be the same event but the way we choose to talk about it seems to be different across languages. This is why for Slobin (1996a), any event (in our case, a motion event) can be described in terms of two different cognitive frames. On the one hand, that which refers to the actual event or experience that we want to describe (the translational motion from one place to another), and on the other, the tools provided to and constraints imposed on speakers in expressing that event in a particular language. What he calls a ‘discourse frame’ and a ‘typological frame’, respectively. If Slobin’s hypothesis is right, then, these differences within the typological frame have to be readily present if we analyse the same type of event in different languages. As we said in the previous section, satellite-framed languages and verb- 10 framed languages differ in the way they lexicalise the motion semantic components of path and manner. S-languages 7 encode some change of location in a particular manner leaving it to satellites (particles and prepositions) to encode directionality, whereas Vlanguages do exactly the opposite. But, how can we empirically show their differences? In his 1996a paper, Slobin proposes a series of ‘hints’ or areas where these differences among S-languages and V-languages become clear. I have organised them in terms of three proposals. Let us define each of them: VERBS: number, expressiveness, and frequency of mention of lexical items for manner description. PROPOSAL 1: V-languages have fewer items than S-languages PHRASES & JOURNEYS: Elaboration of ground descriptions. PROPOSAL 2: V-languages have less frequent and elaborated ground descriptions than S-languages. RHETORICAL STYLE: description of motion vs. scene setting PROPOSAL 3: V-languages devote less narrative attention to dynamics of movement and more to scene setting than S-languages In the following section, I will provide a detailed description of each of these proposals based on both Slobin’s own work and results from English and Spanish 11 (Slobin 1996a, 1997) and my own results from Basque (Ibarretxe-Antuñano 2002, forthcoming). The results of such a description will be discussed in the last section of this paper with respect to the validity of language typologies. 4. Testing Slobin’s proposals in English, Spanish, and Basque frog stories. 4.1. Proposal 1: Verbs. The first step that Slobin takes in his analysis of the typological frame of motion events is to look at the entire collection of motion verbs, both self and caused movement, used in the frog stories. (6a) English verbs [S-96] Buck, bump, buzz, carry, chase, climb, come, crawl, creep, depart, drop, dump, escape, fall, float, fly, follow, get, go, head, hide, hop, jump, know, land, leave, limp, make-fall, move, plummet, pop, push, race, rush, run, slip, splash, splat, sneak, swim, swoop, take, throw, tip, tumble, walk, wander (6b) Spanish verbs [S-96] Acercarse ‘approach’, alcanzar ‘reach’, arrojar ‘throw’, bajar(se) ‘descend’, caer(se) ‘fall’, correr ‘run’, dar-un-empujón ‘push’, dar-un-salto ‘jump’, entrar ‘enter’, escapar ‘escape’, hacer caer ‘make fall’, huir ‘flee’, ir(se) ‘go’, llegar ‘arrive’, llevar(se) ‘carry’, marchar(se) ‘go’, meterse ‘insert oneself’, nadar ‘swim’, perseguir ‘chase’, ponerse ‘put oneself’, regresar ‘return’, sacarse ‘remove oneself, exit’, salir ‘exit’, saltar ‘jump’, subir(se) ‘ascend’, 12 tirar ‘throw’, traspasar ‘go over’, venir ‘come’, volar(se) ‘fly’, volver(se) ‘return’. (6c) Basque verbs Abiatu ‘set off’, agertu ‘appear, turn up’, ailegatu ‘arrive’, airatu ‘fly’, alde egin ‘leave’, aldendu ‘leave, move away’, altxatu ‘raise, get up’, astindu ‘shake’, ateatu ‘go out the door’, atera ‘go/take out’, atzera egin ‘go backwards’, aurreratu ‘go forward’, azaldu ‘turn up, appear’, bidali ‘send off’, bota ‘throw’, bueltatu ‘return’, desagertu ‘disappear’, eragin ‘cause to do something, move’, erakarri ‘cause to bring’, eraman ‘carry’, erori ‘fall’, eskapatu ‘flee’, eskumarantz egin ‘go towards the right’, eten egin ‘stop’, etorri ‘come’, frenatu ‘brake, stop’, gelditu ‘stop’ (remain), gora egin ‘go up’, heldu ‘arrive’, hurbildu ‘approach’, ibili ‘walk’, igo ‘go up’, ihes egin ‘escape’, inguratu ‘go around, get close’, iritsi ‘arrive’, irriste egin ‘slide, slid’, irten ‘go/take out’, itzuli ‘return’, jarraitu ‘follow to’, jarri ‘put in’, jausi ‘fall’, jo ‘set off, head’, joan ‘go’, kanpora egin ‘go outside’, kendu ‘remove’, korrika egin ‘run’, lurreratu ‘go down’, makurtu ‘bend’, montatu ‘mount’, mugitu ‘move’, oheratu ‘go to bed’, pasatu ‘go beyond, pass’, paseatu ‘stroll’, saltatu ‘jump’, salto egin ‘jump’, sartu ‘enter/put inside’, segitu ‘follow’, zutitu ‘stand up’. If we compare the total number of the verbs listed under (6), we see that Basque has the highest number of verb types with 58, followed by English with 47, and then, Spanish with only 27. But, why are these numbers so different? 13 I think there are two independent but related explanations. On the one hand, the fact that Basque has a much higher number of verb types in comparison with English and Spanish is due to the possibilities and resources that this language has for vocabulary creation and extension. In this list, we can observe the following ones: (i) conflation of path with motion as typically from verb-framed languages, e.g. igo ‘ascend’, sartu ‘enter’ (ii) locative noun and the verb egin ‘make’, e.g. alde egin ‘leave’ (iii) locative noun together with an allative or directional allative case, and a verb like egin ‘make’, e.g. gora egin ‘go up (above-all-make)’, eskuma-rantz egin ‘go right (right-dir.all-make)’ (iv) locative noun with allative and a verbal suffix, e.g. lurre-ra-tu ‘go down (ground-all-suf)’, aurre-ra-tu ‘go forward (front-all-suf) (v) Romance loans, e.g. ailegatu ‘arrive’, bueltatu ‘return’ (vi) Pairs of synonyms, e.g. iritsi and heldu ‘arrive’ 8 The possibility of using these strategies for conveying this type of verbs allows the lexicon to be very rich. For instance, if we wanted to say ‘go out’ in Basque, the lexicon would give us the opportunity to choose among four different possibilities: atera, irten, kanpo-ra egin, and kanpo-ra-tu, plus an English like construction with the locative noun (kanpo ‘outside’) with the allative and the verb joan ‘go’, i.e. kanpo-ra joan ‘go outside’ 9 . This strategy does not only concern path verbs but also manner verbs. For instance, ‘jumping’ has two different verbs in the list above saltatu, and salto egin. In the first case, the verb is formed with a verbal suffix –tu, whereas the other case is 14 created by combining the verb egin ‘make’ with a nominal indicating the kind of action performed (‘jumping’), a complex predicate. If we had to list the verbs not just by the different lexical items that we find in these stories but by their different meaning, the Basque list will contain 41 different items instead of the 58 we mention above. A number lower than that of English verbs but still higher than that of Spanish verbs. According to Slobin, another reason for these differences in the number of verb types can be found in the possibility of motion and manner conflation in the verb in satellite-framed languages. In his own words: It is as if the availability of the combined slot for MOTION and MANNER in S-languages has encouraged speakers to elaborate the entries in this slot. There is no additional “cost” to adding richer manner expressions, since the slot must be filled by some verb or other in order for a syntactically complete sentence to be produced. By contrast, the optional slot for MANNER expression in a V-language has some “cost”, in that it adds an element or phrase to the sentence. Thus it is retained for situations in which manner is truly at issue –because it is unexpected or unusual (Slobin 2000: 113) This statement is supported by the data in (6). There are 31 different types in the English list above, that is, a 65% of the total number. The percentage of manner of motion verbs in both Spanish and Basque is indeed much lower. Spanish has nine types (30%) and Basque eleven types 10 (23%). But the difference between these languages not only lies on the number of verb types, but also on the expressiveness of such verbs. Slobin argues that manner 15 verbs in S-languages are more expressive than those in V-languages. Once again, the data confirm this possibility. Spanish and Basque manner of motion verbs are all firsttier verbs, i.e. neutral and everyday verbs such as running (correr, korrika egin) or jumping (saltar, salto egin), whereas in English, manner of motion verbs are very detailed and described very specific movements (plummet, splat, swoop). In sum, Slobin’s first proposal is supported by the data. S-languages do seem to have a more expressive and higher number of manner of motion verbs with respect to V-languages. 4.2. Proposal 2: Phrases and Journeys: elaboration of ground. The second proposal concerns the elaboration of ground, that is to say, the description of source, medium, milestone, and goal. Slobin proposes that V-languages have less frequent and elaborated ground descriptions than S-languages. In order to test this proposal, Slobin focuses his analysis on ground phrases on the one hand, and on journeys on the other. Let us have a look at these two elements. 4.2.1. Phrases One of the problems when comparing and contrasting English, Spanish and Basque motion events is that what is expressed by one single expression in one language requires or is equivalent to more than one in the other language. This is the case of the Spanish verb caer(se) (and Basque erori). As Slobin notices, the verb caer(se) has two translations in English: fell and fell down. The important factor at issue in this case is not the translation per se, but the fact that English offers two somehow different possibilities. The first one is the use of a ‘bare’ verb, i.e. a verb that “provides no elaboration of path beyond the inherent directionality of the verb 16 itself” (Slobin 1996a: 200). The second one is the use of a motion verb plus additional path information in the satellite. The first possibility is obviously the most common choice in V-languages (or at least in Spanish, as we will show later), the information about the path is already given in the verb so there is no need to add an ‘extra’ particle like down in order to convey this information. This particle is, however, compulsory in S-languages since that is the ‘slot’ reserved for path information. If we were going to test Proposal 2 on the basis of these data, that is, on the basis of bare verb usage, we would find that the descriptions of ground are much richer in S-languages than in V-languages. This is indeed what Slobin finds in the data from downward motion descriptions in the Spanish and English Frog stories 11 . But to some extend, these data are misleading. As Slobin points out, if we had two groups on the basis of bare verb usage, it would imply that verbs with descriptions of path such as down or in the water were considered as belonging to the same group, the non-bare verb group. However, as we can see from these expressions, the information they convey is not the same, whereas down refers to downward movement, in the water gives us more explicit information about the ground (goal). The other difference is that the downward information in down is already expressed in the verb caer(se), while in the water needs to be expressed in a separate expression in Spanish as well –en el agua. In order to solve this problem, Slobin (1996a: 201) proposes to analyse the elaboration of ground in terms of ‘Minus-ground clauses’ and ‘Plus-ground clauses’. The former include bare verbs and verbs with satellites indicating direction of movement. The latter have in addition one or more phrases encoding source and/or 17 goal. Table 1 summarises the percentages of minus- and plus-ground clauses in English, Spanish and Basque. INSERT TABLE 1 AROUND HERE The use of minus-ground clauses in English is very low, only an 18%, in contrast with that of plus-ground clauses, an 82%. Spanish shows a more balanced use, a 37% for minus-ground and a 63% for plus-ground. So far, these figures support Proposal 2. However, if we look at Basque percentages we will see that they do not correspond to what Slobin reports as the general tendency for V-languages. The use of minus-ground clauses in Basque is very low, an 11%; in fact, it is even lower than that of English. The use of plus-ground clauses, on the other hand, is high, an 88%. What is more, the type of motion verb it occurs with does not condition this pervasive description of ground. As Ibarretxe-Antuñano (2002) reports neutral verbs like joan ‘go’, path verbs like jausi ‘fall’, and manner verbs like salto egin ‘jump’ appear with ground descriptions. 4.2.2. Journeys After the analysis of clauses, Slobin examines a more realistic narrative description of motion that includes more than just source and goal. As he argues, narrators in real narratives “need not limit a path description to a single verb and its adjuncts […] [they] may present a series of linked paths or a path with way-stations” (1996a: 202). In order to do so, he proposes a new ‘unit of analysis’, what he calls, a ‘complex path’ or ‘journey’, i.e. an extended path that includes milestones or subgoals, situated in a medium. 18 In the case of English, narrators seem to mention several pieces of information about the ground as illustrated in example (7). (7) He starts running and he tips him off over a cliff into the water [S-96] In (7), there is information about the milestone (over a cliff) and the goal of motion (into the pond). These two pieces of information are attached to one single verb. For V-languages, on the other hand, Slobin reports that there is a tendency to limit the description of ground to one piece of information only, and that examples such as (8), where both source and goal are expressed appended to one verb 12 , are quite rare in Spanish. (8) El perro… hace un movimiento tal que se precipita al suelo, desde la ventana the dog makes one movement such that it.refl plummets to.the ground from the window ‘The dog… makes a movement such that he plummets to the ground, from the window’ Although the one-ground-element-per-verb limitation seems to be true for Spanish, what we find in descriptions of the same scene by Basque narrators contradicts this tendency. Let us look at examples (9) and (10). (9) Bapatean Txuri txakurra leihotik behera joan zan [IA-02] 19 suddenly txuri dog:ABS window:ABL below:ALL go:PERF aux ‘Suddenly, Txuri the dog went down from the window’ (10) danak amildegitikan behera erori zian ibai batera [IA-02] all.ABS cliff:ABL:LOC below:ALL fall:PERF aux river one:ALL ‘All of them fell from the cliff down into the river’ In (9) there are two pieces of information: a source lehiotik ‘from the window’ and a directional goal behera ‘down (below.all)’. In (10) there are three: a source amildegitikan ‘from the cliff’, a directional goal behera ‘down (below.all)’, and another goal ibai batera ‘to a river’. Both examples illustrate the clause-compacting strategy that Slobin mentions for English. This type of sentences are very common and natural in Basque, thus there are not exceptions as it was the case in Spanish. In fact, it has been suggested that the use of source and goal of a translational motion in the same clause as in lehiotik behera and amildegitikan behera is a pervasive tendency in Basque. Ibarretxe-Antuñano (2002, forthcoming) proposes that these sentences are cases of what she calls the ‘Complete Path construction’ 13 . In this context, construction is understood in terms of Fillmore and Kay’s (1996) and Goldberg’s (1996) ‘Construction Grammar’, that is to say, basic units of language that carry meaning independently of the words in the sentence. This author argues that this construction 14 is used in those cases when the path is ‘delimited’, i.e. the location of both source and goal is conceptualised as a fixed point in space, and as a result the distance between them (path) is viewed as a restricted, delimited trajectory 20 between these two fixed points. The focus of a CP construction, however, is not on either the source or the goal, but on the path delimited by them, the full trajectory defined by these two points The CP construction, therefore, seems to add further support to the fact that path elaboration in Basque does not correspond to that of a typical V-language such as Spanish. Slobin’s analysis does not stop here, at the description of complex paths. As he argues, languages may differ in the way they structure complex paths –as we have just seen clause-compacting is common in English and Basque, but rare in Spanish-, but we also need to look at the content of the narration, at what is narrated, and see whether speakers of these languages have the same degree of event granularity, the same degree of detailed description for the same event. In order to test this possibility, Slobin chooses the ‘deer scene’ in the Frog stories (see Appendix for the scene pictures). This is a very rich and complex scene that depicts how boy and dog fall from the cliff. It has six different narrative segments: (i) deer starts to run (ii) deer runs, carrying the boy (iii) deer stops at the cliff (iv) deer throws the boy (off the antlers/down) (v) boy and dog fall (vi) boy and dog land in water 21 The goal of this test is to see whether narrators that do not use a detailed description of ground elements ‘compensate’ it by using other strategies. The fact that a language like Spanish does not use clause-compacting does not mean that it cannot show the same degree of event granularity, this language might use separate clauses but offer the same information in the end. Therefore, we need to see how many of the six segments in the deer scene are mentioned in these languages. The results are summarised in Table 2: INSERT TALBE 2 AROUND HERE In the case of English and Basque almost all the narrators provide three or more segments, a 100% in English and a 93% in Basque. These high percentages contrast with those in Spanish, where only a 75% of the narrators do so. On the basis of these figures we can conclude that: (i) there is a direct correlation between ground description and event granularity. Languages with a high degree of ‘ground-plus clauses’ seem to analyse events in more components than languages with a low degree of ‘ground-plus clauses’; 15 (ii) the low use of ground clauses is not compensated “by means of a series of separate action clauses that analyse a journey into its components” (Slobin 1996a: 203-4). 4.3. Proposal 3: Rhetorical style 22 The last proposal deals with the rhetorical style that speakers from these languages employ in the narration of these events. As Slobin points out, despite possible differences in the higher or lower degree of elaboration and description of the ground, English, Basque and Spanish narrations ‘tell the same story’. Slobin argues that this can be explained on the basis of how much narrative attention speakers devote to movement and setting. He proposes that S-languages allocate more attention to the description of movement rather than to that of the physical setting in which the action takes place. This is due to the rich means that this type of languages has to describe path, to the “availability of verbs of motion (often conflated with manner) that can readily be associated with satellites and locative prepositional phrases to trace out detailed paths in relation to ground elements” (Slobin 1996a: 205). On the other hand, V-languages constrained by their typological characteristics take the opposite choice: they leave the path to be inferred and focus on the description of the setting. Let us look at examples (11), and (12). (10) The deer stops abruptly, which causes the boy to lose his balance and fall with the dog down into the stream [S-96] (12) caen en la laguna […] que estaba debajo de ese precipicio [A20a] 16 fall in the pond that was below of that cliff ‘They fall into the pond, which was below that cliff’ In the English example (11), as predicted by Slobin, the location of the river under the place where the deer stops, i.e. the cliff, is inferred by the trajectory 23 described in fall…down into the stream. 17 This becomes very evident if we compare (11) with the Spanish description of the same event in (12). Here, the speaker explicitly describes the position of the pond with respect to the cliff estaba debajo de ese precipicio ‘under that cliff’. But what happens in Basque? As a V-language, Basque should also focus on the physical description of the setting rather than on the dynamics of movement, but if we take into account that the description of path is more similar to S-languages, then we could expect it to behave like English. In fact, this is what we find in the data. An 80% of Basque speakers follow the S-language strategy, that is, the setting information –the location of the river below the cliff- is inferred in the description of the trajectory, adjoined to the verb of motion –from the cliff fall down into the river- as illustrated in (13). (13) danak amildegitikan behera erori zian ibai batera [IA-02] all.ABS cliff:ABL:LOC below:ALL fall:PERF aux river one:ALL ‘All of them fell from the cliff down into the river’ However, we also find examples with a combination of both static information about the physical setting and dynamic information about the trajectory as in (14). (14) eta gure Andoni eta txakurra amildegitik behera erori ziren, baina amildegiaren azpian erreka zegoen zorionez eta ez zitzaien ezer gertatu [B20I] 24 and our andoni and our dog:ABS cliff:ABL:LOC below:ALL fall:PERF aux fall:PERF but cliff:POSS below:LOC river:ABS was fortunately and neg aux nothing happen ‘And our Andoni [boy] and the dog fell down from the cliff, but fortunately, there was a river under the cliff, and nothing happened to them’ Although both techniques are combined in this example, it is important to point out that the static description does not occur on its own as it was the Spanish case. It comes after the complete path description, amildegitik behera ‘down from the cliff’, which already presupposes where the river is located. 5. So now, what we do with these contrastive data? A few conclusions… One of the main goals of this paper was the comparison of three different languages, English, Spanish, and Basque, with respect to their lexicalisation of motion events. Following Talmy’s definition of a ‘characteristic expression of motion’, English was classified as an S-language, and Spanish and Basque as V-languages. As we saw in the examples analysed in Section 1, English expresses manner and motion conflated in the verbs fly and tumble and path in the satellite out. Spanish and Basque, on the other hand, convey path and motion conflated in the verbs salir, irten and atera ‘exit’ and manner in a different expression, a gerund corriendo ‘running’ and an adverb hegaka ‘flying’. However, in order to prove that these languages belong to one group or the other we need more empirical data, more evidence that supports this classification. When we talk about motion we do not just utter one single sentence, a sentence that 25 includes all the key features of this typology, that is, path and manner. In a normal everyday situation, we talk about motion within a discourse context, in an environment where we do not need to mention everything. It is only in this way that we can really see how languages deal with motion. In order to do so, we have analysed motion events from the Frog Stories in these three languages using Slobin’s work in this area. This author argues that an experience cannot be verbalised without having taken a specific perspective influenced, if not determined, by the typological characteristics and lexicalisation patterns of a given language. In other words, the typological characteristics of S- and V-languages drive speakers of these languages to talk about motion in a certain way. In order to test this ‘Thinking for speaking’ hypothesis, Slobin gives us a number of areas where these typological characteristics may show up. We organised them in terms of three different proposals. The results are summarised in Table 3. INSERT TABLE 3 AROUND HERE With respect to the first proposal –number, expressiveness, and frequency of mention of lexical items for manner description- these languages behave as predicted. English, an S-language, has more items than Spanish and Basque, both V-languages. Unfortunately this is as good as it gets; the results from the rest of the proposals do not correspond to the predictions. In general, we could say that Basque does not behave as a typical V-language like Spanish in the elaboration and description of path. Basque fails all the predictions related to this motion component for V-languages. In fact, if we closely look at the results in Table 3, Basque seems to be more similar to English, that is, to an Slanguage rather than to Spanish with respect to the elaboration of path. 26 Proposal two predicted that V-languages had less frequent and elaborated ground description than S-languages. This is indeed what we have for English and Spanish. English rarely uses a motion verb with no elaboration of ground, and what is more, there is usually more than just one ground piece of information per motion event. Spanish, on the other hand, prefers verbs with no elaboration of ground, and in case the verb does have some ground description, this is usually restricted to one piece of information. But when we look at Basque, the situation is strikingly different. Instead of complying with the results from Spanish, Basque goes the opposite way. Verbs do appear with plus-ground clauses; in fact, the percentages are even more dramatic than those in English are. There is only an 11% of minus-ground verb occurrences. The description of ground is not restricted to one piece of information either. The use of source and goal with the same verb is not rare as in Spanish, but very common and natural as the Complete Path construction attests. Further support that confirms that Basque is very sensitive to path description comes from the analysis of event granularity in the deer scene. Out of the six segments that form this scene a 93% of Basque speakers mention more than three of them. This percentage coincides with that of English but not with that of Spanish (75%). Finally, results from Proposal 3 –rhetorical style in these narratives- once again situate Basque closer to English than Spanish. According to this proposal, Slanguages tend to focus more on the dynamics of movement than on the physical description of the setting, the technique followed by V-languages. All of Basque narrators describe the path trajectory in more or less detail, whereas the physical setting is only mentioned occasionally and usually after the trajectory has already been established. This shows that Basque speakers are more sensitive to the dynamics of movement than Spanish speakers are at least in these narratives. However, the fact 27 that some of them offer some static description for the same scenes makes us wonder whether this is just an exception to the rule or common practice. This is a very important question that needs further research because it might have important consequences for Slobin’s third proposal. If the static description is not just an exception, it will mean that Basque speakers focus their narrative attention not only on the dynamics of movement but also on the static description of the setting. Slobin offers us a choice between one strategy or the other, but what the Basque data seem to indicate is that both strategies are alright, one (dynamic) is preferred over the other (static), but both of them are used to some extent. In sum, English, Basque and Spanish behave as predicted by Slobin with respect to the semantic component of Manner (Proposal 1) but not in relation with the semantic component of Path (Proposals 2, 3, and 4). Although Basque and Spanish are both V-languages, the former shows a strong and pervasive tendency to elaborate and describe the Path more often and in much more detail than the latter. After having revised Slobin’s proposals in relation to these three languages, the last question that we need to answer is the following: what consequences do these data bring to Talmy’s two-way typology? Strictly speaking, Slobin’s proposals and the results that we have obtained from this contrastive analysis would not necessarily affect the validity of Talmy’s binary typology. This is based on the way these languages lexicalise the core information of a specific domain, in our case the Path component, and not on how detailed these components are elaborated in narrative discourse. However, the fact that languages can vary so much within the same typological group is an issue that we cannot ignore, that we cannot take for granted for two main reasons. Firstly, it shows us the possible shortcomings of a broad 28 typological classification such as Talmy’s to accommodate existing intra-typological variation. Secondly, it raises a fundamental question about the distinction between the uses of a typology to contrast languages versus accounting for the ways in which particular semantic domains receive expression in connected discourse. In conclusion, the results from this contrastive analysis of motion events in English, Spanish, and Basque do not cast doubt on Talmy’s typology; rather, they add nuances –especially with regard to applications of the typology to analyses of discourse- that should be taken into account if we want to offer a full description of how motion events are expressed both cross-linguistically and within a language. References Ameka, F.K. and J. Essegbey, In press, “Serialising languages: satellite-framed, verbframed or neither”, In L. Hyman and I. Maddieson (eds.), African Comparative and Historical Linguistics (Proceedings of the 32th Annual Conference on African Linguistics, University of California, Berkeley, March 2001), Lawrenceville, NJ, Africa World Press. Berman, R. A., & D. I. Slobin, 1994, Relating events in narrative: A crosslinguistic, developmental study, Hillsdale, NJ, Lawrence Erlbaum Associates. Boas, F., 1911, “Introduction”, Handbook of American Indian Languages, Bulletin 40, Part I, Bureau of American Ethnology, Washington, DC, Government Printing Office. (Reprinted in F. Boas. 1966, Introduction to Handbook of American Indian Languages, Edited by P. Holder, Lincoln, University of Nebraska Press). 29 Etxepare, E., In press, “Valency and argument structure in the Basque verb”, In J.I. Hualde and J. Ortiz de Urbina (eds.), A Grammar of Basque, Amsterdam and Philadelphia, John Benjamins. Fillmore, C. and P. Kay, 1995, Construction Grammar, Stanford, CSLI Publications. Garai, K. and I. Ibarretxe-Antuñano, 2002, “From X to Y: the ‘Complete Path’ construction in Basque”, Odense Working Papers in Language and Communication 23: 289-311. Goldberg, A., 1995, Constructions. A Construction Grammar Approach to Argument Structure, Chicago, Chicago University Press. Hamano, S., 1998, The Sound-Symbolic System of Japanese, Stanford, CLSI Publications. Hinton, L., Nichols, J., & J. Ohala, 1994, Sound Symbolism, Cambridge, Cambridge University Press. Ibarretxe-Antuñano, I., 2002, “Motion events in Basque narratives”, In S. Stromqvist, and L. Verhoeven (eds.), Relating Events in Narrative: Typological and contextual perspectives, New Jersey, Lawrence Erlbaum. Ibarretxe-Antuñano, I., forthcoming, “Basque: Verb-framed or Satellite-framed?”, Linguistic Typology. Ibarretxe-Antuñano, I., In preparation, Tipi-tapa, Tipi-tapa Korrika!!! Motion and Sound Symbolism in Basque. Mayer, M., 1969, Frog, where are you?, New York, Dial Press. Myhill, J., 1992, Typological discourse analysis: Quantitative approaches to the study of linguistic function, Oxford, Blackwell. 30 Sapir, E., [1924] 1958, Selected Writings of Edward Sapir in Language, Culture, and Personality, ed. by D.G. Mandelbaum, Berkeley, Los Angeles, University of California Press. Slobin, D. I., 1987, “Thinking for Speaking”, Proceedings of the Thirteenth Meeting of the Berkeley Linguistic Society, Berkeley, Berkeley Linguistics Society, 335-345. Slobin, D. I., 1991, “Learning to think for speaking: Native language, cognition, and rhetorical style”, Pragmatics 1, 7-26. Slobin, D. I., 1996a, “From ‘thought and language’ to ‘thinking for speaking’”, In J. J. Gumperz and S. C. Levinson (eds.) Rethinking Linguistic Relativity, Cambridge, Cambridge University Press, 195-217. Slobin, D. I., 1996b, “Two ways to travel: Verbs of motion in English and Spanish”, In M. Shibatani & S. A. Thompson (Eds.), Essays in semantics, Oxford, Oxford University Press, 195-317. Slobin, D. I., 1997, “Mind, code, and text”, In J. Bybee, J. Haiman, & S. A. Thompson (Eds.), Essays on language function and language type: Dedicated to T. Givón, Amsterdam/Phildaelphia, John Benjamins, 437-467. Slobin, D. I., 2000, “Verbalized events: A dynamic approach to linguistic relativity and determinism”, In S. Niemeier & R. Dirven (Eds.), Evidence for linguistic relativity, Berlin, Mouton de Gruyter, 107-138. Talmy, L., 1985, “Lexicalization patterns: Semantic structure in lexical forms”, In T. Shopen (Ed.), Language typology and semantic description. Vol. 3: Grammatical categories and the lexicon. Cambridge, Cambridge University Press, 36-149. 31 Talmy, L., 1991, “Path to realization: A typology of event conflation”, Proceedings of the Seventeenth Annual Meeting of the Berkeley Linguistics Society, 480-519. Talmy, L., 2000, Toward a Cognitive Semantics, Cambridge, MA, MIT Press. Voeltz, F.K.E. and C. Kilian-Hatz, 2001, Ideophones, Amsterdam and Philadelphia, John Benjamins. Whorf, B. L. [1940] 1956, Language, Thought, and Reality: Selected Writings of Benjamin Lee Whorf, ed. by J. B. Carroll. New York, Technology Press of M.I.T. and John Wiley & Sons. Zlatev, J. and P. Yangklang, 2002, “A third way to travel: The place of Thai (and other serial verb languages) in motion event typology”, In S. Stromqvist, and L. Verhoeven (eds.) Relating Events in Narrative: Typological and contextual perspectives, New Jersey, Lawrence Erlbaum. 32 90 88 82 80 70 63 60 50 % 37 40 Minus-ground Plus-ground 30 20 18 11 10 0 English Spanish Basque Frog Stories Table 1: Minus- and plus-ground clauses in English, Spanish, and Basque. 33 Language 3 or more segments English 100% Spanish 75% Basque 93% Table 2: Event granularity in the ‘deer’ scene 34 Proposals 1. VERBS total manner 2a. PHRASES minus-ground plus-ground 2b. JOURNEYS complex path event granularity < 3 4. RHETORICAL STYLE setting path ENGLISH = 47 types 65% = 18% 82% = several 100% = inferred described SPANISH = 30 types 30% = 37% 63% = one 75% = described inferred BASQUE = 63 types 23% ≠ 11% 88% ≠ several 93% ≠ usually inferred described Table 3: Results from the comparison of English, Spanish and Basque motion events 35 Appendix 1: The ‘Deer scene’ in Frog where are you? (Mayer 1969) 36 * This research is supported by Grant BFI01.429.E from the Basque Country Government's Department of Education, Universities, and Research. I would also like to acknowledge the support and help I received from the International Computer Science Institute, the Institute of Human Development at UC Berkeley, and the Max Planck Institute for Psycholinguistics, Nijmegen. Part of the research reported here was carried out while I was a research fellow in the Department of Linguistics at the University of California, Berkeley. 1 Examples taken from these studies are always marked with a reference in square brackets. References are abbreviated in the following way: author’s initials and year of publication, e.g. [S-96] stands for Slobin 1996a, [B-S-94] for Berman and Slobin 1994, [IA-02] for Ibarretxe-Antuñano 2002. 2 A ‘satellite’ is defined as “the grammatical category of any constituent other than a noun-phrase or prepositional phrase complement that is in a sister relation to the verb root. [It] can be either a bound affix or a free word” (Talmy, 2000: 102). 3 For a more detailed discussion of the characteristic expression of motion in English and Spanish, see Talmy (1985, 1991, 2000) and in Basque see Ibarretxe-Antuñano (2002, forthcoming). 4 It might be argued that the protagonist’s inner state is not strictly a type of Manner of motion since it only reflects the psychological state or attitude of the Figure (person that moves) and can be also applied to other domains apart from that of motion. However, we think that it is a key factor in the characterisation of this type of events. In the examples that we discuss in this section, all what we have is a set of different adverbials or adjectives that lexicalise these inner states, but we need to take into account that some motion verbs do include this component in their semantics. See for instance, the English verbs wander or strut, or the Spanish deambular ‘saunter’. These verbs, apart from other Manner of motion details, include information about the protagonist’s inner state in their semantics – wander and deambular tell us that the Figure is walking without any special purpose, not only that, deambular also suggests that the Figure is absentminded; strut refers to walking but in a stiff, selfsatisfied way. 5 The use of sound symbolic expressions for describing manner of motion seems to be a common practice in languages with sound symbolic systems (see Hamano 1998, Hinton et al. 1994, Voeltz and Kilian-Hatz 2001). Basque is among those languages and thus, there are several sound symbolic 37 expressions (taka taka ‘small steps’, tirriki tarraka ‘dragging’) and words of sound symbolic origin (irristatu ‘slide’) in this area (Ibarretxe-Antuñano 2002, in prep). 6 Modified because its main goal is not to prove the effects of grammar on worldview or nonlinguistic behaviour, but to show the way in which speakers of a language organise their thinking in accordance to the linguistic tools offered by their native language. 7 From now onwards we will follow Slobin’s abbreviations for satellite-framed languages (S- languages) and verb-framed languages (V-languages). 8 As Etxepare (in press, ch. 4.1) points out it is interesting to notice that some of these pairs of verbs are only synonyms if they are used as intransitive predicates. If we use them transitively, these are no longer synonyms, iritsi would mean ‘reach’, and heldu ‘grab’ (ailegatu would not accept the transitive construction). 9 The reader should bear in mind that we are exclusively referring here to the intransitive physical meaning of these verbs, i.e. go out from a physical place, and not to other possible metaphorical extensions that these verbs may convey as well. In most contexts, these exiting verbs can function as synonyms when describing an exiting event but, of course, as it happens in all cases of synonymy, their equivalence is restricted and constrained by different dialectal and contextual factors. 10 Nine if we only count those with a different meaning (21%). 11 The percentage of bare verbs in English adult narratives is a 15% while in Spanish adult narratives is a 63%. These figures are even more dramatic in children. English pre-schoolers offer a 16% versus a 56% of their Spanish colleagues; English nine-year-olds, a 13% vs. Spanish ones, a 54%. 12 Although Slobin does not mention this issue, I think it is important to point out the structure and order of these two ground elements. First of all, there is a pause (,) that breaks into two units the description, and secondly, the order of these elements is somehow ‘inverted’: the goal of motion is mention before the source of motion. Despite the fact that there is only one verb in this sentence, I think the two elements do not form a conceptual unit as the English example (or the Basque one in (7)), but two units, the second with an elliptical verb. 13 The CP construction is defined as “the tendency to linguistically express in the same clause both the source and the goal of a translational motion, even in cases where one of the components is pleonastic” (Ibarretxe-Antuñano 2002) 38 14 As shown in Garai and Ibarretxe-Antuñano (2002), this construction is not restricted to physical motion contexts. It is used in non-physical motion, and even metaphorical, cases such as argitik ilunera (light.abl dark.all) ‘all day’, hitzetik hortzera (word.abl tooth.all) ‘suddenly’. 15 Data from Thai (Zlatev and Yangklang 2002) and Ewe (Ameka and Essegbey in press) further supports this hypothesis. 16 In Slobin (1996a) the examples that illustrate this point are all drawn from the children’s data. Since we are only discussing the Frog stories in relation with adult data I have taken this example from the adult speakers in the same Spanish corpus compiled by Aurora Bocaz in Argentina. I like to thank Dan Slobin for giving me access to these data. 17 This is even clearer in example (5) where the speaker says over the cliff into the river describing the whole trajectory from source to goal. 39