1. Introduction

2Although the semantic description of posture verbs is neither unknown nor necessarily rare cross-linguistically, their semantic properties have never been previously explored in to details in any of the Gur languages as far as I am aware. A sketch of the posture phenomenon is to be found in my own previous and recent works (Atintono 2004b, 2009, 2011b), and Brindle & Atintono (2011). In Atintono (2004b), I present a brief analysis of the semantics of the three posture verbs gã ‘be lying’, zĩ ‘be sitting’, ze’ ‘be standing’ in Gurenɛ in the light of Talmy’s (1985; 2000a; 2000b) model of motion events (i.e., the general category of MOTION includes motion and static location). The paper sketches the use of the posture verbs to describe human postures and the entities that can occur in the basic locative construction as Figure or Ground elements without any detailed discussion. Atintono (2009) on the other hand, examines the verbs that can occur in the basic locative construction (BLC) and its structure in line with Ameka & Levinson (2007a) typology in a paper presented at the Sixth World Congress on African Linguistics (WOCAL6) in Cologne, Germany. Similarly, my recent work, Atintono (2011b) is a paper that I presented at the 42nd Annual Congress on African Linguistics (ACAL42), University of Maryland, US. It discusses the aspectual properties of the positional verbs pointing out relevant linguistic tests that can be used to determine the stative and the dynamic positional verbs in the language. Brindle & Atintono (2011, in press) is also a proceedings paper that compares topological relation markers in Gurenɛ and Chakali (a Gur language in Ghana). None of these works presents any comprehensive analysis of the posture phenomenon.

3In this present paper, I examine the meanings of the posture verbs largely from a descriptive and cognitive linguistics perspective drawing on concepts such as image schemas, domains, and conceptualizations to explore fully the meanings of these verbs taking into account the linguistic and cultural systems that contribute to their interpretations. Apart from the basic meanings of the three posture verbs, which describe the ‘sitting’, ‘lying’, and ‘standing’ postures of humans, two of these verbs (gã ‘be lying’, ze’ ‘be standing’) are also extended to describe the location of other inanimate entities. In this respect, pots, cups, bottles, buildings, poles, tables, clothes, and balls, located in space can be described by the speaker as standing or lying. The use of the posture verbs to describe the location of entities other than humans provides a window on how Gurenɛ speakers conceptualise and categorise entities located in space.

4This study, is also motivated by recent typological studies on the posture and positional verbs, which suggest that there is a diversity in the way speakers use posture verbs to describe the location of objects which reflects their cognition of space (Levinson 1992; Lee 2001; Newman 2002a, 2002b; Levinson & Wilkins 2006a; Hickman & Robert 2006; Ameka & Levinson 2007a; Evans & Levinson 2009). Consequently, to understand how speakers think about and describe space, a detailed investigation into the semantics and grammar of the posture verbs in a particular language can be particularly insightful (cf. Schaefer & Egbokhare 2008: 215). In line with this typology, I show in this paper how Gurenɛ speakers use posture verbs to describe and configure this positional space.

5The present discussion draws on a number of works in the typological literature on posture verbs (see Lemmens 2001, 2002; Newman 2002a and the contributions therein; Hellwig 2003; Levinson & Wilkins 2006a; Ameka & Levinson 2007a; Lemmens & Perrez 2010). In particular, a significant portion of my discussion of the basic and extended uses of the Gurenɛ posture verbs follows Lemmens (2002). Drawing on concepts from cognitive linguistics like prototypes and image schemas, Lemmens presents a detailed account of the semantic network of Dutch posture verbs (staan ‘stand’, zitten ‘sit’, and liggen ‘lie’) by characterizing their basic meanings associated with human postures and their extension to describe the location of other animate and inanimate entities as well as providing some insights into their metaphorical extensions. The systematic way in which he presents his analysis of the meanings of the Dutch posture verbs serves as motivation for my analysis of the Gurenɛ data. As will be shown in the discussion, apart from points of agreements there are also striking differences between Gurenɛ and Dutch (or in English and the other languages in the typology) in the way the speakers conceptualise posture scenes of similar entities (see section 5).

7The paper is organized as follows: section 2 lays out the structure of the locative construction in Gurenɛ while section 3 discusses the theoretical cognitive linguistics background that is adopted for the analysis. Section 4 focuses on the methodology and in section 5 I discuss the basic locative meanings. Section 6 looks at the extended uses. Section 7 presents some concluding remarks.

2. Static locative constructions in Gurenɛ

8This section lays out the basic structure and the elements in the Gurenɛ locative construction. The static locative construction in Gurenɛ (see schema below) consists of the noun phrase (NP) in the subject position which represents the Figure (the located entity), a posture verb describing the configuration of the Figure followed by a focus particle, a postpositional phrase (POSTP) containing the NP expressing the Ground (the reference point) and a body-part postposition which specifies the search domain (i.e., the precise place of location of the Figure on the Ground). The expressions in the examples below represent typical locative expressions in Gurenɛ. They refer to statements that speakers provide as a natural response to questions demanding the location or whereabouts of people and objects or spontaneous statements that speakers make about the location of entities in a discourse context without a question posed to solicit for the response. The locative statements which represent responses to elicitation questions as in examples (1)-(2) and similar data in the rest of the paper correspond to the basic locative construction (BLC) in the sense of the MPI research tradition on space proposed by Levinson & Wilkins (2006a) and Ameka & Levinson (2007a). The BLC refers to the canonical (unhesitated, spontaneous, non-elliptical) and the typically preferred answer to the question where is an entity x? The elicited examples involving the use of the various stimuli represent the BLC type of responses but the naturally occuring data on locative descriptions in everyday contexts (e.g, homes, pubs), like the spontaneous data (see example (3)) do not involve a question posed to elicit this kind of response similar to Newman’s (2002a) typological approach). In Gurenɛ stative locative descriptions there is no other alternative for describing location than to use the posture predicates and the postpositions. However, it is perfectly possible to describe locations within view of the participants, within a radius of 15m, by using deitic locative adverbials in a verbless clause as attested in (4) such expressions are usually accompanied by pointing to draw attention to the location. They are, however, used in a very restricted context and are generally not considered the most preferred way for describing locations, since they provide minimal or no information about the Figure or Ground. The four examples below are taken from three different types of datasets (the Gur Positional Photos stimulus-GUR, the Spontaneous Speech Text-SPST, and the Locative Description Finding Task-LDFT). The details of all the data types are discussed in section 4. The spatial locative construction in Gurenɛ can be schematized as stated below.

Schema : NP (Figure) V(Posture verb) FOC NP(Ground) POSTP(LOC)

(1)	Bia	la	gã	la	suŋɔ	la	puan
	child	DEF	lie.STAT	FOC	mat	DEF	inside
	‘The child is lying in (on) the mat.’ (GUR 05)

(2)	Pɔka	la	zĩ	la	kuka	la	zuo
	woman	DEF	sit.STAT	FOC	chair	DEF	head
	‘The woman is sitting on the chair.’ (LDFT 52)

(3)	Wanɛ	la	n	ze’	la	tiŋa	dikɛ	kabegɛ
	calabash	DEF	which	stand.STAT	DEF	land	take	fetch
	Ko’om	la	bo	saana	la	la
	water	DEF	give	visitor	DEF	DEF
	‘Take the calabash which is standing on the ground and fetch water for the visitor. ‘(SPST 75)

(4)	E	bala	bilam
	3SG	DEF	there
	‘S/he is located over there.’

9The posture verbs code stative (STAT) locative relations as shown in these examples. That is to say, the verbs express the static location of the Figures (cf. Newman 2002b on English posture verbs). The Figure elements in the examples are represented by the NPs: bia ‘child’ in (1) and pɔka ‘woman’ in (2). These are animate human Figures whose postures are described by the posture predicates gã ‘be lying’, and zĩ ‘be sitting’ respectively. Example (3) on the other hand, wanɛ ‘calabash’ is an inanimate Figure with its location described by ze’ ‘be standing’ through extension because it is supported on its base and projects vertically (see details in section 6 on locative extensions). The Ground elements in these examples are expressed by the NPs suŋɔ ‘mat’ in (1), kuka ‘chair’ in example (2), and tiŋa ‘land, (i.e., ground or earth’) in (3). Observe that Gurenɛ uses mostly body-part terms as postpositions to express points of reference for the spatial orientation of objects, and these are grammaticalised with their semantics corresponding to typological findings on body-part terms in African and other languages (see Heine et al. 1991: 152‑153; Levinson 1994: 801; Svorou 1994 on body-part terms and spatial relations). That is, the primary references of these terms designate names for human body-parts but are recruited as grammatical concepts for spatial reference. Body-part terms in Gurenɛ as in other languages (see Ameka 1995 on Ewe) constitute a form class of nouns with their own morphology and semantic properties. The search-domain that specifies the exact region where the Figure is located on the Ground is marked by the postpositions puan ‘inside, in (1) and zuo ‘on top, upper surface’ in (2) respectively. Puan is a grammaticalised reflex of the body-part term puurɛ ‘stomach’, the postposition zuo is a body-part noun referring to head. The third postposition, tiŋa ‘ground, floor literally land or earth’ (see example (3)) is not a body-part term. Whenever the latter is used it inherently includes location and this explains why the search-domain element represented by a postposition can be omitted.

10An interesting culture-specific point to note about the use of the postposition puan in (1) is that in Gurenɛ, suŋɔ ‘mat’ is woven of long guinea grass stalks and can be rolled up in a bundle (similar to a carpet). It is about three to four metres long. People spread out part of the mat on the floor to lie on, and roll the other part to cover themselves when the weather is cold. The use of the mat in this way lets speakers perceive it as a containing Ground that contains people even when they are not using it to cover themselves. Hence, the use of puan ‘inside’ instead of zuo ‘on top, upper surface’. Usually zuo is used when someone is lying on a bed or a modern plastic mat which cannot be rolled for use as a cover. The use of puan in this context is an image schematic representation of the mat as a container. That is, it reveals the recurrent use of puan by speakers to show their experience of using the mat to cover, categorised in this specific context as a schematic instance of a container.

11One other important element in Gurenɛ locative constructions is the particle la which occurs as a definite article after nouns but as a discourse particle marking focus in postverbal position (see examples (1)-(3)). The particle la, in Gurenɛ and in most of the other Gur languages, occurs in different syntactic positions to express different grammatical functions (see Bodomo 1997 on Dagaare; Issah 2007 on Dagbani, Dakubu 2000 on Gurenɛ). Its multiple functions in Gurenɛ include focus, definite article, NP linker, gender-neutral 3rd person singular, and a demonstrative. The phonological shape of the particle is the same in all the positions where it occurs. I do not intend to review all these functions of the particle in this paper because its other roles are not of any immediate relevance in this current discussion. However, in the locative construction, its precise functions are restricted to the definite article and focus marking. As a focus particle, it conveys emphatic communicational value on the verb and the postpositional phrase that follows it in the locative construction.

3. Theoretical background

13In cognitive semantics, meaning reflects conceptualization as shaped by linguistic conventions (Langacker 1987; Clausner & Croft 1999). The postural phenomenon which is of interest in this paper is linked to spatial concepts which are based on our experience with our evironment. The domain of experience relevant to this paper is that of spatial relations. Our everyday interaction and observation of the world leads to emergent conceptual representation known as image schemas or image-schemata which are abstracted from our sensory and perceptual experience as we manipulate objects, position ourselves spatially and temporally, and direct our perceptual focus for various purposes (Evans & Green 2006: 178‑182; Lee 2001).

14Newman (2002: 2) identifies four cognitive domains that are relevant for the interpretation of three posture verbs in English (sit, stand, lie): spatio-temporal domain, the force-dynamic domain, the active zone domain, and the socio-cultural domain. The spatio-temporal domain relates to the overall spatial configurations associated with each posture. That is, the posture encoded by each of the postural verbs indicates the position of the Figure maintained through time. The Force dynamic domain, on the other hand, characterizes the manner in which entities exercise or balance force to be in a particular posture (see also Talmy 2000a: 409‑422). Drawing on Langacker (1991: 189‑201), Newman uses the term active zone to refer to the specific area or subpart of an entity that participates directly in a spatial relation. For the posture verbs, there is typically one part of the body that is more engaged than another (e.g., the legs and feet for standing). The socio-cultural domain refers to the worldview of the speakers and how they conceptualize the various postural states. Some of these domains are relevant for my analysis such as the spatio-temporal and force-dynamic domain. For example, sitting and lying postures are generally construed as comfortable positions for rest and other activities that do not require physical movement, for example in a meeting or sleeping (see Song 2002). Among Gurenɛ speakers, people who are engaged in jobs that require extensive sitting such as office work are not considered to be doing any valuable work. They describe people engaged in such jobs as being lazy people who do not want to sweat on the fields but sit in one place performing no action. Thus, the force-dynamic domain aspect that speakers associate with sitting jobs requiring no physical effort leads to the social evaluation of such jobs.

15In cognitive semantics, the conceptualization principle explains that there is no isomorphic mapping of elements of the external world into linguistic form instead, every situation can be construed in different ways and the different ways of presenting a situation represent different conceptualizations (Lee 2001: 2). This principle forms the basis for my discussion of the meanings of the three Gurenɛ posture verbs and their extensions. For instance, the extension of the basic meanings of the human posture verbs to describe the location of inanimate objects can best be explained in terms of how the speakers conceptualise these objects in terms of human postures. Crucial concepts here are the horizontality and verticality that are associated with these postures. Further, cognitive semantics views meaning as encyclopedic which refers to our comprehensive knowledge about the meaning of a concept (Langacker 1987: 4.2.1).

16My analysis of the Gurenɛ posture verbs follows the view that these are prototype categories, where their basic meanings concern the three human postures while the extensions are considered to be more peripheral.

4. Methodology

18My own fieldwork data include, (i) the Gur Positional Verb Photos (GUR) that I designed consisting of 72 picture scenes depicting various cultural-specific postures and locations in the Gurenɛ community and (ii) the Locative Description Finding Tasks (LDFT) involving elicitation using real objects arranged in different positions in context for speakers to describe. I also collected Spontaneous Speech data (SPST) based on overheard speech at different discourse settings (e.g., homes, pubs, market). The Interactive Discourse Text (IDT) is yet another natural discourse data source which consists of the observation of speakers talking about or describing the actual location of objects in the community (e.g., settlements, or items displayed in the market). The final pool of text data that is included is based on recorded and transcribed texts of folktale genres (FT) from my documentation corpus. In all ten native-speaker consultants described the picture stimuli during the elicitation sessions. Each consultant described a set of stimuli in an individual session before participating in a group session where the differences noted in the individual sessions were discussed to seek explanations for the variations.

19The total number of posture and locative expressions collected from these diverse set of data in the corpus for the three posture verbs is 956, distributed as follows: 157 tokens for zĩ, 359 for ze’ and 440 for gã. Given that not all the recorded data have been transcribed these figures remain modest; unlike many European languages, Gurenɛ does not have any organised written corpus that one can rely on such as the British National Corpus (BNC). The sit verb zĩ shows up infrequently in the data mainly because its use is restricted to human posture; it is used to refer to animals (e.g, rabbits, rats, leopards, elephants, goats, sheep, etc.) but only rarely, in the limited context of the folktale genre where these animals are characters in the tales and take on human qualities.

5. The basic meanings of the posture verbs

20The basic meanings of the posture verbs describe the human postures: gã ‘be in a lying posture’, zĩ ‘be in a sitting posture’, and ze’ ‘be in a standing posture’. These basic meanings constitute the prototypical uses of the posture verbs. That is they represent examples that are easily called to mind, and frequently used to describe human postures. They are the most typical cases of the use of these verbs to describe human postures. In this sense, the description of human postures becomes the default interpretation that speakers associate with the meanings of these verbs (see Lemmens 2002: 104‑105 on Dutch; Serra Borneto 1996: 460 on German). The picture scenes below depict the three prototypical human postures described by the three posture verbs, gã ‘be lying’, zĩ ‘be sitting’, and ze’ ‘be standing’.

21Language internal evidence that can be used to support the claim of the basic usage of the posture verbs in Gurenɛ is that when referring to human posture, they have a wider application as compared to their use to describe the location of other objects. For example, gã can be applied to both a person lying on a mat and on a bed (see (5), (6)). However, the same does not hold for balls lying on the ground or on the bed as examples (7) and (8) show. Instead, the more specific positional verb dɔgi, see example (9) is to be used because the balls are objects with an unstable base located on an elevated ground. Recall that verbs of elevation are to be used for the description of location on elevated Grounds.

(5)	Bia	la	gã	la	suŋɔ	la	puan
	child	DEF	lie.STAT	FOC	mat	DEF	inside
	‘The child is lying in (on) the mat.’ (GUR 05)

(6)	Bia	la	gã	la	gɔregɔ	la	zuo
	child	DEF	lie.STAT	FOC	bed	DEF	head
	‘The child is lying on the bed.’

(7)	Bɔɔla	la	gã	la	tiŋa.
	ball.PL	DEF	lie.STAT	FOC	land
	‘The balls are lying on the ground.’ (PSPV 39)

(8)	*Bɔɔla	la	gã	la	gɔregɔ	zuo
	ball.PL	DEF	lie.STAT	FOC	bed	head
	‘*The balls are lying on the bed.’ (PSPV 39)

(9)	Bɔɔla	la	dɔg-i	la	gɔregɔ	la	zuo
	ball.PL	DEF	be on top.unstable.base-STAT	FOC	bed	DEF	head
	‘The balls are on top (of unstable base support) of the bed.’

22The descriptions of the three different postures in the scenes above depict the physical body position or orientation of the people. Thus, gã ‘be lying’ describes the child’s horizontal body position (GUR 05) which is in contact with the mat (see example (5)) while example (10) zĩ describes the woman’s sitting posture (it depicts the sitting posture that is typical for women in the culture). Example (11) on the other hand, ze’ ‘be standing’ focuses on the vertical orientation of the man on his feet on the ground.

(10)	Pɔka	la	zĩ	la	suŋɔ	la	puan
	woman	DEF	sit.STAT	FOC	mat	DEF	inside
	‘The woman is sitting on (in) the mat.’ (GUR 10)

(11)	Budaa	la	ze’	la	tiŋa
	man	DEF	stand.STAT	FOC	land
	‘The man is standing on the ground.’ (GUR 48)

23From these basic usages, the meaning of these verbs is extended to describe the location of non-human entities (both animate and inanimate). Before turning to these extended uses, the next section describes the basic uses of each of the verbs in more detail.

5.1 gã ‘be lying’

24The posture verb gã ‘be lying’ typically describes the horizontal orientation of animate Figures (humans and animals). Our discussion will be limited to human posture. The verb describes a person whose whole body or part of the body is in contact with and aligned with a horizontal Ground, such as the scene of the child lying on a straw mat in Figure 1 above (GUR 05). The lying postures may refer to a person lying on one side of the body, on the back with face up or lying prostrate on one’s stomach. However, the latter two postures would sometimes attract the use of two adverbials, aliko ‘prostrate’, and azampuyɛla ‘on the back facing up’ to indicate the precise body position of the person lying. In this type of construction, the Ground element is usually omitted as well as the search domain. What is of interest or in focus to the speaker in these two contexts is the manner of the lying posture.

(12)	Bia	la	gã	la	aliko
	child	DEF	lie.STAT	FOC	prostrate
	‘The child is lying prostrate.’

(13)	Bia	la	gã	la	azampuyɛla
	child	DEF	lie.STAT	FOC	on.back
	‘The child is lying on his back.’

25The verb is also used to describe someone lying on a bed. The response in (15) is taken from my dataset of spontaneous speech. The utterance was provided by a woman in response to my consultant’s question (see (14)) about the whereabout of her (my consultant’s) child who was playing in a yard with other children while we were having an elicitation session. The child had gone to lie on the bed without our knowledge.

(14)	N	bia	la	boi	la		bɛ?
	1POSS	child	DEF	exist	FOC		where
	‘Where is my child?’
(15)	A	gã	la	gɔregɔ	la	zuo
	3SG	lie.STAT	FOC	bed	DEF	head
	‘He is lying on the bed.’ (SPST 67)

26In Gurenɛ, horizontality is an important feature for defining the lying posture of humans and animals but not objects, as we will see later (see section 6.2). Ameka (2007) also points out that in Likpe (a Kwa language spoken in Ghana) inanimate Figures such as pens, sticks, tubers, bottles, that have an elongated shape and are clearly in a horizontal position on a surface (e.g., table, floor) are never described as lying but are categorised as in a ‘be.on’ relation. According to Ameka, the contexts in which the lying verb can be applied to objects in Likpe are restricted. For example, when there are multiple objects in different orientations and Likpe speakers intend to show contrast in the positions of the multiple Figures (e.g., three bottles standing and four bottles lying on a table), the lying verb is used in this context (see Likpe example in 16).

(16)	ə-tsywə́	nyə	a-na	labe	lə́	ɔ-punu	əsúə́
	AGR-three	stand	AGR-four	lie	LOC	CM-table	surface
	‘Three are standing, four are lying on top of the table.’ (Ameka 2007:1090)

27The important semantic property that accounts for the non-application of the lying verb to objects in Likpe “as Ameka observes” is animacy. In Gurenɛ, however, animacy is not an issue as far as the extended use of the lying verb is concerned. Rather it has to do with whether the object has a canonical part to support it on its base or lacks a salient dimension (e.g., if it is round). Kutscher and Schultze-Berndt (2007: 999) note that in colloquial German locative descriptions, objects lacking salient dimensions are also described as lying. If a Figure does not have base support in Gurenɛ then it will receive gã ‘be lying’ by default. In Gurenɛ, holes dug in the ground are described as ‘lying’. It can be tentatively argued that the fact that horizontality is not an important construal of the Gur and Kwa languages suggests an areal typological phenomenon. However, such a conclusion requires further verification with data from other languages in the linguistic area.

28The Gurenɛ lying verb gã also shows a difference with other languages in that it is not applied to people lying on a high Ground especially when the Ground is not a canonical place for lying. For example, a person lying on a bed is described as gã but when they are in a lying posture on a tree branch or on the roof of a bus (a common practice in the community), they are described with pagi ‘be on top, of flat or flexible objects’ to indicate that they are on a higher Ground. That is because the latter two Grounds (branch and bus top) are not perceived by the speakers as appropriate or natural places for someone to be lying on. In the community, there are plastered roofs which are specially designed for people to climb up and rest or sleep on (see Figure 5 below, GUR 12) when the weather becomes hot at certain times of the year (usually March-May). When people are in a lying posture on this type of roof they could be described as lying because these roofs are made especially for this purpose and are thus considered as canonical. However, if the person is lying on a roof made of thatch roof or corrugated iron sheets, they will never be described with gã ‘be lying’ but with pagi. The issue of elevation applies here just as it does with the zĩ ‘be sitting’ and ze’ ‘be standing’ discussed below (section 5.2 and 5.3).

29As in other languages (see Ameka 2007: 1090), gã has another sense associated with ‘sleeping’, metonymically associated with lying. However, these two senses can always be differentiated by the adjuncts in the locative construction. The locative sense always includes the Ground element without the verb gisa ‘sleeping’ added (see example (17)) while the other sense of gã meaning ‘sleep’ often occurs as the first verb in a serial verb construction and usually collocates with the verb gisa ‘sleeping’ as in example (18) and (19).

(17)	Bia	la	gã	la	gɔregɔ	la	zuo
	child	DEF	Lie.STAT	DEF	bed	DEF	head
	‘The child is lying on the bed.’

(18)	Bia	la	gã	gisa	mɛ
	child	DEF	lie.STAT	sleep.IPFV	FOC
	‘The child is lying and sleeping.’

(19)	Bia	la	gã	la	gɔregɔ	zuo	gisa
	child	DEF	lie.STAT	FOC	bed	head	Sleep.IPFV
	‘The child is lying and sleeping.’

5.2 zĩ ‘be sitting’

30The posture verb zĩ is used to describe the sitting posture of humans located on their buttocks with support from below. This is the only posture verb that is restricted to the description of human postures. It has no extension to the description of the location of inanimates. Typical sitting postures described by speakers are those represented in Figure 2 below taken from the Gur positional photos (see GUR 01, GUR 06, GUR 07) and one scene from the topological relations picture series (TRPS 38).

31Prototypical sitting postures in Gurenɛ require that the person’s buttocks are supported on the Ground with legs stretched out (like the woman in GUR 07) or bent (see the postures of the two children and the man in Figure 2).

32One interesting cultural-specific conceptualization associated with the sitting posture with legs bent, like the man’s sitting posture or the children (GUR 01, 06, TRPS 38) is that it marks respect or deference for authority. This is the type of posture assumed by contestants for a chieftaincy title or complainants and defendants who appear before their traditional chiefs in their palaces. The photo in Figure 3 below shows a real contextualised example of contestants for a chieftaincy title at one of the palaces in the Gurenɛ speaking community (Bongo). As observed during the fieldwork, failure to assume such a posture before the chief is considered a serious offence and disrespect for authority. Thus, the image-schema of this sitting posture in Gurenɛ profiles a spatial relationship that marks high and low status distinctions between the authority holding power and the ordinary subjects with no power. This aspect of the meaning cannot be thought of as pertaining to the meaning of zĩ ‘be sitting’ independently of certain contextual assumptions, including culture-specific background assumptions against which the inferences can be assessed appropriately. These inferences are however, not built into the linguistic structure of the sentences that give rise to them (cf. Levinson 1983: 167). Gurenɛ, unlike Chitimacha (Amerindian) where Newman (2002b: 20‑21) citing Swadesh (1966: 322) reports that the language uses three different auxiliary morphemes (ci(h) ‘standing’, pe(h) ‘lying’, and hi(h) ‘neutral with respect to posture’) to indicate different kinds of honorific notions Gurenɛ does this through pragmatic inference from the body posture. For instance, Newman points out that if the lying auxiliary occurs in a sentence in Chitimacha it ascribes the notions of insult, sarcasm, disparagement, joking, abuse, and defiance, while the standing auxiliary shows respect. In Gurenɛ, sitting with the body tilted or remaining in a standing posture when you should be seated constitutes an insult or protest.

33

34In Gurenɛ, sitting on a chair, a log or any other object also attracts the use of zĩ but only when the feet are supported on the ground (floor or earth). Similarly, sitting with legs crossed is expressed with the verb zĩ provided one of the legs is in contact with the ground. In these two situations, if the person is seated with his feet or legs suspended or not supported on the Ground (e.g., sitting on a high chair or a stool behind tills at shops) the speakers never use zĩ. Instead, a different verb yagi ‘be on top, with base or stable support’ is applied to indicate that the person is suspended on an elevated ground. Recall that the semantics of gã ‘be lying’ discussed above shows some restrictions when the Figure is construed as located on a non-canonical high Ground. Zĩ ‘be sitting’ in contrast does not have these restrictions of a canonical high Ground. The important semantic factor is that the person’s feet are suspended from the Ground.

35This shows that Gurenɛ zĩ unlike its equivalent in other languages (see for example Lemmens 2002 on Dutch) has an orientational commitment to the posture of humans. Lemmens (2002: 105) points out that the absence of orientational commitment with the Dutch posture verb zitten ‘sit’ triggers its use to describe different body positions that include sitting on the buttocks, a chair, resting on the knees, on hands and knees, and on all fours. Gurenɛ is different from this and will use different specific positional verbs to describe all these varied body positions. For instance, resting on the knees will attract kpa ‘kneel’, and resting on hands and knees will be predicated with nyigi ‘kneel stooping’.

36The projection of the body vertically while in a sitting posture is also crucial for Gurenɛ speakers. Thus, if the person is not oriented vertically but in a tilted or inclined posture, as illustrated in Figure 4 below, a different posture verb deli ‘lean, sitting or reclining’ will apply to schematise the posture as one that is not vertical but in a reclining position.

37The corresponding expression that speakers use to describe this posture is given in (20) below. However, sometimes when speakers are pressed further to find out whether they would never apply the sitting verb zĩ to describe such postures they offer a serial verb construction to say that the person is sitting and leaning as example (21) illustrates.

(20)	Budaa	la	del-i	la	kugedɛleŋa	la	zuo
	man	DEF	lean-STAT	FOC	chair.tilt	DEF	head
	‘The man is leaning on the reclining chair.’ (GUR 21)

(21)	Budaa	la	zĩ	dɛl-a	la	kugedɛleŋa	la
	man	DEF	sit.STAT	lean-STAT	FOC	chair.tilt	DEF
	‘The man is sitting leaning on the reclining chair.’ (GUR 21)

38The importance of verticality for the (postural) use of zĩ in Gurenɛ thus clearly distinguishes it from that of the equivalents in other languages that has been described in the literature such as in English or Dutch (Lemmens, personal communication).

39Another feature in which Gurenɛ deviates from other languages is that zĩ is not applicable to a person sitting straddling a tree branch, a wall, or a bike carrier or even sitting on a wall or tree branch, which are at a certain elevation where the person’s feet are not touching the ground. Consider Figure 5, below with the various stimuli pictures involving sitting straddling postures (SUP 29, SUP 31, GUR 03, and GUR 11). The corresponding expressions are given below.

(22)	Budaa	la	yag-i	la	tia	yile	zuo
	man	DEF	be on top-STAT	FOC	tree	branch	head
	‘The man is on top of (with base support) the branch of the tree.’ (SUP 29, 30)

(23)	Bia	la	yag-i	la	tia	yile	zuo
	child	DEF	be on top-STAT	FOC	tree	branch	head
	‘The child is on top of (with base support) the branch of the tree.’ (GUR 03)

(24)	Budaa	la	yag-i	la	dangoone	la	zuo
	man	DEF	be on top-STAT	FOC	wall	DEF	head
	‘The man is on top of (with base support) the wall.’ (SUP 31, 32)

(25)	Bia	la	yag-i	la	keekee	kariya	la	zuo
	child	DEF	be on top-STAT	FOC	bike	carrier	DEF	head
	‘The child is on top of (with base support) the bike’s carrier.’ (GUR 02)

(26)	Dunkiima	la	yag-i	la	boŋa	la	zuo
	shepherds	DEF	be on top-STAT	FOC	donkey	DEF	head
	‘The shepherds are on top of (with base support) the donkey.’ (GUR 11)

(27)	Pɔka	la	yag-i	la	gɔse-wanɛ	la	zuo
	woman	DEF	be on top-STAT	FOC	roof-calabash	DEF	head
	‘The woman is on top of (with base support) the edge of the wall.’ (GUR 12)

40Like gã discussed in 5.1 above, these straddling sitting postures are considered by speakers to be located on elevated grounds and in this case, what takes precedence is the elevation of the location, which then serves to override the actual posture or orientation of the Figure. In this case, yagi ‘be on top, with base support’ will be the default choice by speakers. That is because speakers construe these straddling postures as location on a high ground. However, if the wall or stool is very short permitting sitting astride with the feet of the person touching the ground (earth) speakers agree that zĩ can be used in this context. The relevant semantic feature for the use of the verb is for the person sitting to have his feet touching and in contact with the surface of the earth while remaining in a sitting posture. As pointed out before, Gurenɛ has a set of what I call “verbs of elevation” which speakers select to describe elevated locations thus disgarding the Figure’s actual posture. This set of elevation predicates includes dɔgi ‘be on top, of unstable base support of the Figure’ (e.g., ball on tabletop), pagi ‘be on top, of flexible or spread object (e.g., cloth on a tabletop), and yagi ‘be on top, with base support of the Figure’ (e.g., cup on a table or the straddling examples above). This class of verbs is used specifically to predicate this type of location according to the Figure’s properties (e.g., shape, position, configuration). This feature is where Gurenɛ shows deviation from the other languages in the typology (see for example Ameka 2007: 1086 on Likpe; Lemmens 2001, 2002 on Dutch) where sitting or straddling positions are described using the sitting posture verbs in these languages.

41The verb zĩ is the only posture verb among the three that is restricted to the predication of human posture. It is not used to describe the posture of animate non-humans (e.g., animals) and the location of inanimate entities. For example, if a scarecrow were made to sit on a chair, yagi ‘be on top, with base support’ would be used. However, in the folktale data, animal characters such as rabbits, monkeys, tigers, and hyenas are sometimes described as sitting when the narrator personifies them with human attributes. Example (29) below describes Mr. Rabbit (who is a main character in Gurenɛ folktales) in one of the tales that I collected in my documentation corpus as sitting with his family in the front yard of his compound when Mr. Leopard arrived to question him about a theft case in his (Mr. Leopard’s) home. This usage is permissible because the animals in the folktale genre are personified and therefore are perceived as behaving like humans. This makes it possible for them to assume the sitting posture in the folktale discourse.

(29)	Asɔ’ɔŋa	la	a	pɔga	la	a	kɔma	zĩ
	Rabbit	CONJ	3SG	wife	DEF	3SG	head	Sit-STAT
	la	ba	yire	zanyɔren	sɔsera	dee	ti	Abaa
	FOC	3PL	house	front	converse	and	COMP	Leopard
	paɛ…
	reach…
	‘Mr. Rabbit and his wife and his children were sitting at the front yard of their house conversing when Mr. Leopard arrived…’ (Folktale 26_20100430)

42This usage in Gurenɛ is also quite different from what is found in other languages where the ‘sit’ verb can often be used to predicate the postures of animals and objects in ordinary discourse or speech. For example, Lemmens (2002: 107) points out that in Dutch when speakers perceive the postures of animals to be sufficiently similar to that of humans they are coded with the verb zitten. Thus, in Dutch, when cats, dogs, cows, have their hindlegs bent and their bottom touching the ground they are described as sitting. Similarly, Ameka (2007: 1086) reports that in Likpe, si the verb for coding a sitting posture applies to cats and dogs resting on their behind. Kutscher & Genɕ (2007: 1046) in their discussion of Laz positional verbs also observe that animals in such postures can be described as sitting. In Gurenɛ, such postures of animals are also described with the verb dɔ̃bi ‘be squatting.’ Examples in the stimuli include scenes of a cat resting on its hindlegs under a table (TRPS 31), a dog in squatting position by a doghouse (TRPS 6), and another dog in a bowl resting on its hindlegs (TRPS 47). Thus, speakers do not consider this posture of animals as sitting and will argue that animals do not sit but squat. Smaller animals such as rats, mouse, and frogs when resting on their hindlegs are described as squatting. Notice that Dutch uses zitten ‘sit’ here as well (Lemmens 2002). It should be pointed out that in Gurenɛ coding something as a sitting posture as shown in the stimuli photos above (Figure 2) requires the legs to be bent or stretched out and these animals postures do not align very well with this requirement.

5.3 ze’ ‘be standing’

43The typical postural meanings of ze’ include the location of a human being on his feet or base, vertical projection from the base, and vertical orientation (cf. Lemmens & Perrez 2010: 318‑319 who make similar observations for Dutch staan). For an illustration, see Figure 1 above (GUR 48). In Gurenɛ, for someone to be described as standing requires the location of the Figure on the earth or at the floor level. Like gã ‘be lying’, and zĩ ‘be sitting’ if the Figure is on a raised or high Ground (such as someone standing on top of the roof like those in Figure 6 below), ze’ would not be eligible for the coding.

44For these picture scenes, speakers used yagi ‘be on top, with base support’ to describe the location of the Figures in a standing posture. The exact expressions that speakers used to describe these scenes are given in examples (30)-(33). Notice that this constraint is indifferent to whether scenes involve artificially elevated Grounds (GUR 09, TRPS 34) or natural elevated Grounds (GUR 04, SUP 28). Invariably, my consultants coded people observed in standing posture on top of elevated Grounds such as hills, mountains, raised platforms, vehicles (pick ups, open trucks) with yagi ‘be on top, with base support’ and not with ze’ ‘be standing’.

(30)	Budaa	la	yag-i	la	gɔsegɔ
	man	DEF	be on top-STAT	FOC	plastered.roof
	‘The man is on top (base support) of the plastered roof.’ (GUR 09)

(31)	Bia	la	yag-i	la	tia	zuo
	child	DEF	be on top-STAT	FOC	tree	head
	‘The child is on top (with base support) of the tree.’ (GUR 04)

(32)	Budaa	la	yag-i	la	tia	zuo
	man	DEF	be on top-STAT	FOC	tree	head
	‘The man is on top (with base support) of the tree.’ (GUR SUP 28)

(33)	Nɛra	la	yag-i	la	deo	la	zuo
	man	DEF	be on top-STAT	FOC	room	DEF	head
	‘The man is on top (with base support) of the room/roof.’ (TRPS 34)

45Although speakers did not deny the upward extension of the people in the scenes above from the base (which is crucial for the use of the verb ze’ ‘be standing’ for humans) the elevation of the Ground on which they are resting appears to demote the idea of verticality. Thus, the elevation of the Ground is a central semantic property that can override the use of the posture verb ze’ in these locative contexts as was noted for gã ‘be lying’ and zĩ ‘be sitting’ in 5.1 and 5.2 respectively.

46An important point to note here is that when the speaker who is conceptualising the locative scene is also located on top at the same level as any of those Figures in the scenes in Figure 6 above yagi is still preferred compared to ze’. However, if for some pragmatic reasons consultants want to differentiate two or more people located on top of the same roof to say that some are sitting and others are standing, then they may use zĩ or ze’ to differentiate them in this restricted context.

47Although, elevation has been reported in some studies on the typology of posture verbs, there are differences between Gurenɛ and these other languages. For example, Song’s (2002: 365‑366) work on Korean posture verbs reports that a crucial semantic factor underlying the use of the Korean verb se ‘stand’ to describe other animate or inanimate entities relates to the relative height of a given entity to that of humans. In other words, the vertical length of any entity that is comparable to a human height or higher can access se but if this is lower the verb can no longer be used. It is inferable from Song’s discussion that the phenomenon has nothing to do with location at a certain elevation, as is the case with Gurenɛ.

48Similarly, it is also well reported in the positional verb typology that the Mayan languages (Yucatec, Mopan, Tzotzil, and Tzeltal) and in particular Tzeltal, have a relatively large set of positional verb roots that can be used to provide fine-grained semantic descriptions of all kinds of locative situations between the Figure and Ground including location on elevated Grounds (see Levinson & Haviland 1994; Brown 1994, 2006; Bohnemeyer & Brown 2007). However, many of the semantic properties in Tzeltal are Figure-centred with virtually nothing on the Ground or the nature of the spatial relation. For example, Brown (1994: 760‑761) reports that in Tzeltal, the location on a tabletop of animate entities attracts mochol ‘be-located of an animate lying curled on its side (e.g, cat lying on a tabletop)’ and tek’el ‘be-located, of vertically standing (on hind legs) animate creature or long/thin object (e.g., toy man standing on a tabletop)’. Tzeltal also has a semantically neutral preposition ta meaning ‘at’ which sometimes occurs with the dispositional roots (see Brown 2006: 241). In her discussion of the Tzeltal phenomena, it is suggested that elevation does not play any major role as information on the elevation of the Figures was neither mentioned nor commented on but left implicit. As Grinevald (2006: 42‑43) also notes, the semantics of Tzeltal positionals characterise different kinds of information about the Figure such as shape, texture, size, disposition but the topological relation is left to be inferred.

49Another language in the posture verb literature where a brief comment is made about the phenomenon of elevation of the Ground is Trumai (a Brazilian isolate). Guirardello-Damian (2002: 164‑167, 2007: 937‑938) observes that in Trumai, the semantics of the posture verb tsula ‘lie’ overlaps with another posture verb chumuchu ‘lie’ but tsula is applied to Figures located on Grounds above floor level with chumuchu designated for location at floor level. She explains that a shirt on the floor will attract the use of chumuchu but tsula is grammatically not acceptable in this context. Instead, a shirt on a tabletop tsula can apply. Given these patterns of usage, one might assume that there is a clear division of labour between the two verbs. However, Guirardello-Damian (2002: 166) points out that, it is when the Figure is inanimate that the difference between the two verbs becomes more marked, less so when the Figure is animate. In the latter case, the choice between either verb becomes a matter of pragmatics depending on whether the speaker wants to focus on the lying posture itself or the Figure’s location above the floor level. A clear case of elevation where tsula is used but chumuchu is denied, as Guirardello-Damian shows in his discussion, is when human beings or animals are lying in a hammock.

50Although the use of tsula appears to be similar to one of the Gurenɛ verbs of elevation pagi ‘be on top, of flat or flexible objects’ used for Figures in a horizontal position on elevated Grounds, there are differences in many respects. First, tsula refers to the elevation of the Ground but with no information about the Figure’s property but Gurenɛ verbs of elevation do provide such information. A further point that suggests Trumai’s tsula is not exclusively used for location above floor level in Guirardello-Damian’s discussion is that tsula can in fact be applied to describe a sock on a foot because it (foot) is both horizontal and vertical. It is therefore hard to point to any feature of elevation in this case since the foot is assumed to be at the floor level. Further, in Trumai, examples of scenes with a ball on a tabletop, a stick lying on a tabletop, and a glass on its side on a bench are all described with chumuchu but not tsula (see Guirardello-Damian 2002: 162), suggesting that speakers use these verbs freely. In all these instances, the use of the posture verb is not overridden as it is the case in Gurenɛ. Thus, the Gurenɛ phenomenon deviates from Trumai’s system as well as from Tzeltal and Korean.

51One other important feature about the Gurenɛ data concerns the canonical orientation of the Figure for ze’ to be applicable. For example, if the person is not standing on his feet but is supported on the Ground on his hands or head with his body projected in an upright position he is not described as standing, since he is not resting on his feet. Instead, he would be described as located inversely on the Ground with the expression in (34). Lemmens (2002) points out that for Dutch such an inverse posture is typically described with staan ‘stand’ (with added modifier, meaning ‘on its head’). Apparently, the issue of verticality plays a role in Dutch but not in Gurenɛ in this context. Instead, canonical orientation on the feet is crucial in Gurenɛ.

(34)	A	dikɛ	la	a	zuo	tulege	kpa	tiŋa
	3SG	take	FOC	his	head	turn.over	be.fix	land
	‘He turned his head over and put the head on the Ground.’

52What makes the Gurenɛ case unique is that the elevation becomes an important semantic feature that affects the way the standing and sitting posture verbs can be applied; in addition, for ze’ to be applicable, it is important that the person is oriented with his feet down.

6. Extended locative meanings

53The extended use of the posture verbs to refer to the location of non-human animate entities and inanimate entities is restricted to gã (‘be lying’) and ze’ (‘be standing’); zĩ (‘be sitting’) is never used to refer to the location of non-human entities. The following two sections thus only talk about the former two verbs. The verb zĩ can be used in metaphorical expressions involving humans which are, however, not considered in this paper.

6.1 Location of animate non-humans

6.1.1 gã ‘be lying’

54The verb gã ‘be lying’ is also used to localise the lying postures of animate non-humans such as animals, birds and reptiles as the following examples show. It can be used to describe the lying postures of animals like cats, dogs, cows when they are lying curled on one side of their body on the ground (earth).

(35)	Baa	la	pue	gã	la	bɔ’ɔ	la	nuurɛ
	dog	DEF	be.cross	lie.STAT	FOC	room	DEF	mouth
	‘The dog is lying across the entrance of the room.’ (LDFT 43).

(36)	Bunsɛla	la	gã	la	muɔ	la	puan	ti
	snake	DEF	lie.STAT	FOC	bush	DEF	inside	COMP
	baa	nyɛ	e
	dog	see	it
	‘The snake is lying in the bush and the dog saw it.’ (SPST 82)

55In example (35), my consultants described the posture of a dog that went to lie down at the entrance of our room after we had driven it out because it was interrupting our elicitation session. In (36) a dog saw a snake in a thicket near the place where we usually gather for our folktale narration sessions and one speaker reported this to us. Notice that in all these examples the posture of these animate entities is construed as oriented horizontally on one side of their bodies on the Ground which is similar to human postures. The elongation of their bodies is also an important component that contributes to defining their lying posture. This suggests a transfer of the speaker’s conceptualization of the lying posture of humans to these entities based on the arguments put forward in support of the basic meanings of these verbs (see section 5). An important component of characterizing animals as lying in Gurenɛ therefore is the alignment of the Figure’s body on the Ground (earth). When birds also rest their bodies on the Ground like chickens roosting, they are described with gã. Insects in Gurenɛ are also coded as lying although they are supported on their legs.

(37)	Bɔgereŋa	la	gã	la	tiŋa
	black.ant	DEF	lie.STAT	FOC	land
	‘The black ant is lying on the ground.’

56Speakers construe the legs of ants not to be high enough to project them vertically in a standing posture as compared to other animals (cf. below).

57The issue of elevation of the Ground also affects the description of the lying posture that animals assume. For instance, animals in a lying posture located on a fairly high Ground such as in the support stimulus (SUP 27) where a leopard is lying on a tree branch (see Figure 7 below) ze’ was not used. Rather, a different verb pagi ‘be on top, of flat or flexible objects ’ selected from the verbs of elevation was used to describe this scene.

58In this particular example, because speakers construed the leopard lying on its stomach as being on its flat side on the branch, they applied pagi. In this context, focal advantage or perspective is an important construal influencing their choice of verb. While in Gurenɛ it is not possible to use gã ‘be lying’ in this context, in Likpe in contrast, as Ameka (2007: 1090) observes, the posture verb si ‘be lying’ can still be used to localise scenes like a tiger on a tree branch.

6.1.2 ze’ (‘standing’)

59The standing postures of animate non-humans (animals), require the Figure to be vertically projected on its feet or base and be able to exercise some force in maintaining the standing posture. This characterization of the meaning of ze’ is extended to describe the location of animate non-human entities that are capable of maintaining an upright position. Thus, animals (e.g., cattle or birds) that have legs and can stand are described as standing.

(38)	Nii	la	ze’	la	bu’ɔ	puan	ɔbe-ra
	cow.PL	DEF	stand.STAT	FOC	valley	inside	chew-IPFV
	'The cattle are standing in the valley grazing.' (SPST 69)

(39)	Dayene	la	ze’	la	tiŋa	di-ta
	dove	DEF	stand.STAT	FOC	land	eat-IPFV
	‘The dove is standing on the ground eating grains.’ (SPST 70)

60In example (38), the speaker described a herd of cattle grazing in a valley as standing. Observe that the postposition puan ‘inside’ is also used in this example because Gurenɛ speakers conceptualise any natural geographic feature that is low or has the tendency to collect water such as valleys, rivers, streams, and gullies as containers when humans or objects are located in them. All kinds of birds (domestic and wild, small and large) are described as standing once they have legs, as illustrated in (39). Gurenɛ is thus different from Dutch where for smaller birds (except large ones like ostriches and flamingos) and small animals zitten (‘sit’) is used, cancelling out the fact that they are on their feet (Lemmens 2002: 107). Thus, in Dutch, their overall shape like a crouched posture of a human is of importance. Gurenɛ on the other hand focuses on the legs of these animals which project them vertically. Even smaller animals like mice, rats, black moles, are all characterised as standing. However, animals like hedgehogs whose legs are not quite visible or insects such as ants are described as lying because speakers often perceive them as not having legs.

61Like humans, when birds are perched on trees, walls, ropes (its ends tied to opposing supports), or vehicle or rooftops, they are described as yagi ‘be ontop, with base support’.

(40)	Silega	la	yag-i	la	tia	zuo
	hawk	DEF	be on top-STAT	FOC	tree	head
	‘The hawk is on top (base support) of the tree.’

(41)	Buasi	la	yag-i	la	tia	la	tile
	goat.PL	DEF	be on top-STAT	FOC	tree	DEF	trunk
	‘The goats are on top (with base support) of the tree trunk.’ (IDT 241)

62Again this is because they are located on an elevated Ground and supported on their base, as in (40). Similarly, when goats climb up, and stand on the trunk of low lying trees or tree stumps, yagi is used to predicate their location as attested in (41).

6.2 Location of inanimate entities

6.2.1 gã ‘lying’

63The objects that can attract a coding with gã ‘lying’ include sticks, stalks, cloths, pens, books, mats, wood, (elongated) food items (e.g., yams, cassava), and brooms which assume a lying position. Consider the following examples taken from real context and spontaneous speech data that describe objects in lying positions.

(42)	Sɔɔ	la	gã	la	tiŋa
	broom	DEF	lie.STAT	FOC	land
	‘The broom is lying on the floor.’ (LDFT 18)

(43)	Suŋɔ	la	gã	la	bɔ’ɔ	la	puan
	mat	DEF	lie.STAT	FOC	room	DEF	inside
	‘The mat is lying in the room.’ (LDFT 07)

(44)	Dibega	la	gã	la	tiŋa
	stick	DEF	lie.STAT	FOC	land
	‘The stick is lying on the ground.’ (LDFT 71)

(45)	Daam	dɔɔrɔ	la	gã	la	deo	la	sia
	beer	wood	DEF	lie.STAT	FOC	room	DEF	waist
	‘The firewood for brewing local beer is lying at the foot (lit. waist) of the building.’ (SPST 06).

64As can be seen in these examples, most of the entities described as lying are located horizontally on the Ground. It would appear that the crucial determining factor is that the shape of the Figure must be elongated. However, speakers also apply the verb to describe the location of a ball on the ground (earth), as the following example attests.

(47)	Boole	la	gã	la	tiŋa
	ball	DEF	lie.STAT	FOC	land
	‘The ball is lying on the ground.’ (PSPV 07)

65As said before, although the ball does not have an elongated shape, the use of the verb to describe this scene is consistent with Serra Borneto’s (1996: 464) and Lemmens’s (2002: 122) observation that objects lacking salient dimensions and a canonical base to support them in a vertical or standing position select ‘lying’. This also appears to be the case in Gurenɛ, where any object lacking salient dimension and is located on the ground (earth) without a canonical base support or dimension receives gã as the default coding. For instance, balls, stones, CDs, seeds, pumpkins, watermelons, eggs (see Figure 8) and many other objects with round or irregular symmetrical shapes are all described as gã ‘be lying’ if located at the floor or ground level. It does not matter whether the object is a single Figure or multiple Figures located on earth, gã ‘be lying’ will apply.

(48)	Gɛla	la	gã	la	tiŋa
	egg.PL	DEF	lie.STAT	FOC	land
	‘The eggs are lying on the floor.’ (GUR 70)

66In line with my earlier observations, gã is not applicable when these objects are located on higher grounds (e.g., tabletops, or platforms). For example, in the stimuli data, scenes such as a ball on a tabletop (PSPV 21), 3 small balls and 3 big balls on a tabletop (PSPV 08), 5 big balls on a tabletop (PSPV 18) and 6 eggs on a tabletop (GUR 71), were all localised with the elevated positional verb dɔgi ‘be on top, with unstable base support’. Recall that the choice of any of the verbs of elevation is based on the relevant schematizations of the elevated scene with respect to the Figure properties that the speaker profiles. In the posture typology, it is argued (see, for example, Lemmens 2002: 122 Dutch), that the relative size could be a determining factor for speakers to perceive bigger objects (e.g., very large inflated balls on a pitch) lacking base support as having a vertical dimension and therefore describe them as standing. Size appears not to be an issue in Gurenɛ as speakers explained that such objects would still be described as gã ‘be lying’. Indeed, a huge rock of about seven metres high was described as gã not withstanding its elevation. In sum, for Gurenɛ speakers the notion of verticality is inferior to the notion of a base support.

67In these examples, the horizontal extension of the Figure is no longer a crucial factor for the determination of the lying posture of these objects. Instead, the absence of any canonical base support of the Figures provides a strong motivation for speakers to describe them as lying (cf. Kutscher & Schultze-Berndt 2007: 999‑1000 on colloquial German positionals).

68Speakers also use gã (‘be lying’) to describe the location of some natural phenomena to include water bodies (e.g., rivers, dams, streams), farmlands, geographical features,roads, or foot paths. These are permanently located entities. The following utterances, were spontaneously collected in context. In (49) one of my folktale narrators describes the location of his farmland to his colleague who enquired about this. The description of the dam in (50) was an utterance in answer to an inquiry about the location of the the Vea dam (not far from Bolga). The context in (51) is that we lost our way while travelling to one of the villages to record folktales and we asked a man for directions and he produced that utterance in response.

(49)	Mam	va’am	la	gã	la	kulega	la	nuurɛ
	1POSS	farm land	DEF	lie.STAT	FOC	river	DEF	mouth
	‘My farm land is lying by the edge of the river.’ (SPST 28)
(50)	Vea	mɔgerɛ	la	gã	la	Gu’urɔ	sukuu	la
	Vea	dam	DEF	lie.STAT	FOC	Gu’urɔ	school	DEF
	pooren
	back
	‘The Vea dam is lying at the back (behind) of the Gowrie school.’ (SPST 08)

(51)	Palɛ	la	gã	la	bala	paara	Luŋɔ
	road	DEF	lie.STAT	FOC	DEM	reach	Luŋɔ
	‘The road lies (extends) as far as Luŋɔ.’ (SPST 72).

69In the case of examples (49) and (50), the motivation for the use of gã is in line with what Serra Borneto (1996) calls geotopographical location in his discussion of German. This term ‘geotopographical location’ refers to the conceptualization of a plane as lying. Permanently located landmarks such as water bodies tend to be described as lying in some languages, for example, Trumai (see Guirardello-Damian 2007). Thus, the configuration of va’am‘farmland’ and mɔgerɛ ‘dam’ are conceptualised in terms of their horizontal expansion. Palɛ ‘road’ in (51) on the other hand, is also construed as elongated and thus horizontally extended and hence the motivation for lying. In Dutch, Lemmens and Perrez (2010) observe a similar usage with this type of Figure and even suggest in their findings that French learners of Dutch often have difficulty in understanding this extended usage. Based on this kind of schematizations, an elongated mountain is in Gurenɛ described as lying with gã (cf. also Lemmens on Dutch liggen), but a conical hill which is not extended horizontally is coded with ze’ ‘be standing’. Other geographic entities that are described as lying because of horizontal expansion include forests, unhabited spaces, and sand. However, trees and grasses that grow in the forest are described as ze’ ‘be standing’; these are considered to be vertically projected on their base and are conceptualised as such by speakers as mentioned earlier.

6.2.3 ze’ (‘be standing’)

70When ze' (‘be standing’) is used to describe the location of inanimate objects they may have a base or a leg-like part or any other non-canonical part (e.g., the shorter side of a book) that can allow it to stand upright (cf. Atintono 2004, 2009). The ability of the Figure to maintain a stable upright position requires rigidity of its body. Thus, flexible, non-rigid objects (such as ropes, clothes, or empty sacks) cannot be described as standing since they lack this feature. Instead, they are described as gã ‘lying’. Examples of objects that were characterised with ze’ include pieces of furniture such as tables, chairs, sofas, beds, and books (e.g., a book put to stand on its side). These can all be described as ze’ ‘standing’. Similarly, buildings, trees, cars, bikes, bowls, cups and pots are all conceptualized as standing in Gurenɛ as shown in the examples below.

(52)	Sukuu	la	ze’	la	bɔka	nuure-n
	School	DEF	stand.STAT	FOC	stream	mouth-LOC
	‘The school building is standing by the edge of the stream.’ (SPST 71)

(53)	Teebule	la	ze’	la	bɔ’ɔ	la	puan
	table	DEF	stand.STAT	FOC	room	DEF	inside
	‘The table is standing in the room.’ (LDFT 01)

(54)	Tu’a	la	ze'	la	yire	la	dapoore
	Baobab.tree	DEF	stand.STAT	FOC	house	DEF	back
	'The baobab tree is standing at the back of the house.' (SPST 46)

(55)	Firigi	la	ze’	la	bɔ’ɔ	la	puan
	refrigerator	DEF	stand.STAT	FOC	room	DEF	inside
	‘The fridge is standing in the room.' (LDFT 11)

71Notice that all the objects whose locations are described in these examples have a canonical base support that projects the Figure upright. Again, we see that what is at issue here is an image schema triggering conceptual scanning upward, as Serra Borneto (1996) argues, so that the human posture of standing is transferred to describe the location of non-animate entities.

(56)	Bɔgetɛ	la	ze’	la	bɔ’ɔ	la	tiŋa
	bucket	DEF	stand.STAT	FOC	room	DEF	land
	The bucket is standing on the floor of the room.’ (LDFT 28)

(57)	Bɔgetɛ	la	vugi/kpabi	la	bɔ’ɔ	la	tiŋa
	bucket	DEF	turn.face.down	FOC	room	DEF	land
	‘The bucket is inverted on the floor of the room.’ (LDFT 29)

73Why objects not located on their canonical base cannot be described as standing in Gurenɛ can best be explained in terms of image schematizations. Recall that, as pointed out in section 5.3, when humans are projected on their heads or hands, Gurenɛ speakers do not consider them as standing since they are not located on their feet. Since the inanimate objects are conceptualised in terms of an image schema based on the human posture, it is legitimate to argue that ze’ ‘be standing’ cannot describe the location of inanimate objects in such an inverse position. This is therefore one of the situations where verticality is cancelled out or demoted in favour of a non-canonical locative description. What is still valid here is that if any of these objects is located on an elevated Ground, the choice will fall on one of the verbs of elevation.

7. Conclusion

74This paper has discussed the basic meanings and extended uses of the three posture verbs, gã 'be lying', zĩ 'be sitting', and ze' 'be standing' in Gurenɛ employing concepts from cognitive linguistics for the analysis. It has been shown that the basic meanings of these verbs concern the actual human postures of lying, sitting, and standing respectively. The posture meanings associated with humans are considered basic because their use to describe human postures is what is perceived as the default usage by speakers. In addition, the interpretation of the orientation of other animate non-human and inanimate enties are often interpreted in terms of human postures. The basic uses of gã ‘be lying’, and ze’ ‘be standing’ (but not zĩ ‘be sitting’) extend to describe the location of inanimate entities such as tables, sticks, buildings and natural phenomena such as farmlands, forests, and bodies of water. The extension of two of the posture verbs (gã ‘be lying’ and ze’ ‘be standing’) to characterise the location of other animate (e.g., animals, birds, reptiles, etc.), and inanimate (pots, poles, mats, cups, bottles, bowls, etc.) entities, is also motivated by image schematizations. That is, speakers associate the prototypical standing and lying postures of humans as being in a vertical orientation and projecting from or being in a horizontal position on a Ground. These image schemas are extended to code these other entities located in space. Thus, a bottle is ze’ ‘be standing’ when located on its base but it is described as gã ‘be lying’ when located on its side. The posture verb zĩ ‘be sitting’ is restricted to humans because no animal or any other object is conceptualised by Gurenɛ speakers as assuming such a position. Although, the posture of animals such as cats and dogs when resting on their behinds is close to this sitting posture, Gurenɛ speakers do not conceptualise this as a sitting posture like Dutch or Likpe does (cf. Lemmens 2002, Ameka 2007). Instead, dɔbi ‘be squatting’ is applied by Gurenɛ speakers to describe such postures of animals.

75A unique feature of the Gurenɛ posture verbs is that the location of humans, animals, or entities on an elevated ground leads the speaker to conceptualise the scene from a different perspective which prevents the use of posture verbs, and in Gurenɛ calls for the use of different positional verbs that I call verbs of elevation such as yagi ‘be on top, with base support’, dɔgi ‘be on top, of unstable objects’, and pagi ‘be on top, of flat or flexible objects’. The use of these positional verbs codes the Figures as located on top of the Ground and underspecifies the entity’s actual postures of standing, sitting, and lying. This feature distinguishes Gurenɛ from the other languages described in the literature on posture verbs.

76The theoretical implications for how speakers think about space and orientation is that the Gurenɛ data provides some insights into our understanding of the conceptual packaging of locative descriptions with respect to sensitivity to postural distinctions based on construal, perspective, and image-schemata. Thus, if the speaker construes the location of the Figure to be on the earth the actual posture verb applies, but if the location is construed to be on an elevated ground the actual posture is neutralised, calling for the use of different verbs, notably verbs of elevation in Gurenɛ. This phenomenon in Gurenɛ can contribute towards clarification of the range and type of distinctions to be accounted for in the semantic typology of the posture or positional verbs.