Static locative expressions in Amharic

2.1 Motion event typology

Talmy (2000: 25, 2007: 70) defines Motion event (with capital M) as an event or situation that comprises an object (Figure) moving or located with respect to another object (Ground). It subsumes stationary location and motion (Talmy 2007: 70). The complex Motion event is made up of a framing event (main event) and a subordinate event (co-event) (Talmy 2000: 215).

The main event of Motion can be analyzed into four basic semantic elements: Motion, Path, Figure, and Ground. Figure entity is an entity which is conceptualized as moving or located with respect to another object (cf. Talmy 2000: 25). Ground entity is the entity which is conceptualized as reference entity. Talmy (2014: 3) states that Path is the core schema of the framing event. The Motion component (with a capital M) refers to the occurrence per se of motion or locatedness in the event (Talmy 2000: 25, 2007: 70). The Path (with capital P) component of Motion event is the trajectory that is followed or the site that is occupied by the figure with respect to the ground.

On the other hand, the subordinate event (Co-event) consists of manner, cause, precursion, enablement, and concomitance (Talmy 2000: 27). Note that the Co-event is conceived as support relation to the framing event. Semantically, the Co-event is more primary and vivid than the main event because the framing event is abstract, while the subordinate event tends to be more perceptually palpable (Talmy 2000: 219–20).

Based on the characteristic pattern in which the conceptual structure of the macro-event is mapped onto the syntactic structure, languages of the world can be classified into two categories –verb-framed languages and satellite-framed languages (Talmy 2000: 222, 2007: 153, 2014: 2). Verb-framed languages are languages that characteristically map the core schema of framing event (i.e., Path) into the verb whereas satellite-framed languages characteristically map the core schema onto a satellite and/or a preposition. According to Talmy (2014: 2), the term “satellite” refers to “a constituent in construction with the main verb (root) and syntactically subordinate to it as a dependent to a head.” And the term “preposition” can be defined as “a constituent in construction with an NP that could consist of a preposition and/or postposition and/or affix on the noun of the NP.” Regarding the mapping of the subordinate event, satellite-framed languages map the Co-event into the main verb. On the other hand, verb-framed languages map the Co-event either onto the satellite or into an adjunct (an adpositional phrase or a gerundive type constituent) (Talmy 2000: 222).

However, Huang and Tanangkingsing (2005), Pourcel and Kopecka (2005), and Slobin (2004) a.o., show that some languages do not fit in these typological patterns. Thus, Slobin (2004: 244) added a third motion-framing typology, ‘equipollently-framed category’, in which a language may express path and manner by equivalent grammatical forms. Equipollently-framed languages include: (a) serial-verb languages in which manner and path verbs come consecutively; (b) bipartite- verb languages in which both manner and path are expressed in a verb; (c) preverb languages in which both manner and path may be expressed in the preverbs (see Slobin 2004: 249, 2006: 65).

Although Motion event consists of movement and stationariness (Talmy 2000: 25, 2007: 70), a number of previous studies including those on Ethiopian languages (e.g., Debela 2007; Meyer 2007; Treis 2007; Wakasa 2016) determine the typology of Motion events of a language only based on movement expressions. This makes their claims on Motion event typology of a language incomplete. This is because while some languages like English use the same pattern in both movement and location expressions, other languages such as Amharic differentiate these. For instance, Talmy (2000: 26) takes both English movement and location expressions into account and shows that English employs a satellite-framed Motion construction. Consider the following examples:

1. Motion: The pencil rolled off the table.

2. Location: The pencil lay on the table (Talmy 2000: 26).

In these utterances, Paths (i.e., a path and a site) are encoded by off and on, respectively. And the conflation of Motion and Manner is lexicalized in the verbs rolled and lay.

The present study demonstrates variability of Motion construction patterns between movement and static locative expressions. As we will see in Section 3, Amharic uses a satellite-framed pattern in stationary location constructions exclusively. But it predominately employs a verb-framed pattern in complex clauses of movement expressions as reported in Wakasa (2016: 154).

2.2 Typology of locative predicates

Languages of the world deploy different classes of linguistic units e.g., adpositions, nominal predicates, case inflections, locative verbs when expressing basic location (Levinson and Wilkins 2006b: 16). The term basic location construction refers to the construction that is used in responses to “where – questions” (i.e., “Where is the X?”) (Levinson and Wilkins 2006a: 514, 2006b: 15).

Ameka and Levinson (2007: 863–64) identified four basic types of locative predication based on the size and semantic type of verbs used in the BLC (i.e., type 0, type I, type II and type III). Type 0 includes languages that do not have verbs in a locative statement. Type I refers to languages that have a single locative verb (or suppletion under grammatical conditioning). The locative verb can be (a) a copula or (b) a locative (+existential). Type II comprises languages that employ a small contrastive set of locative verbs (3–7 verbs). The verbs can be (a) postural or (b) ground space denoting verbs. The last type, Type III, includes languages that use a large set of dispositional verbs (9–100 verbs). Ameka and Levinson (2007: 864) assert that genetically, typologically, and areally related languages can belong to different types of locative predication typology.

As will be shown in more detail in Section 4, Amharic employs three types of constructions: (1) a construction with a generic locative verb, (2) a construction with the copula verb nə- ‘be’ and, (3) a construction with an existential copula all- ‘exist’. However, the typical BLC of the language is a construction that involves the verbal copula nə-. In this study, I claim that Amharic belongs to type Ia in the typology of locative predicates developed by Ameka and Levinson (2007).

On the other hand, the preference for the use of positional verbs and copulas or existential expressions can be used to classify languages as positional focus and existence focus. Languages whose speakers tend to use positional verbs are said to be positional focus languages, and those in which speakers frequently employ copula or existential constructions are said to be existence focus (Durst-Anderson 2011 cited in Ng 2014: 338). Since the typical BLC of Amharic is a copular construction, Amharic represents an existence focus language.

2.3 Methods and data collection

The data used in this paper are from the Amharic dialect of Godʒdʒam. They were mainly collected through elicitation and double checked with various speakers. The process was supported by Bowerman and Pederson’s (1992) topological relations picture series. This widely used stimuli kit is prepared to obtain linguistic expressions of spatial relations between figure and ground entities and it includes a one-page instruction sheet. Following the procedure, I showed the picture stimuli to the consultants individually and elicited possible locative descriptions that are natural ways of answering the “where” question about entities in the pictures. In situations where the participants are not able to identify objects in the stimuli, I explained the scenes to the participants. For example, if picture 02 of TRPS is not clear to the participant, I describe the stimulus to the informant as follows: the picture in the yellow color is a spoon and the picture in the black color is a towel. Now, you are going to answer the question “Where is the spoon?” Also, I applied real scenes of the positioning of human figures as stimuli. For the elicitation, fifteen consultants from different Godʒdʒam areas namely, Mot’a, Dəbrəwork’, Dəbrə Mark’os, K’ʷarit, and Yɨsmala have participated. See Table 1 for the demographic information on the consultants.

3.1 Verbs^[2]

Amharic has several positional verbs expressing the conflation of locatedness and manner related to positioning of a figure at a site. Such verbs convey information about the situation of locatedness (lack of motion) and also the posture of a (human) figure at a site. They can be termed manner conflated static posture verbs. Based on the positioning types of the figure, they can be classified into four categories: (a) verbs of standing, (b) verbs of sitting, (c) verbs of lying, and (d) verbs of hanging.^[3] With the exception of verbs of hanging, Amharic positional verbs are mainly used when locating a human figure. In the following sections I briefly discuss the four verb categories one by one.

3.2 Summary

In Section 3 I described the semantics of Amharic static locative expressions by using Talmy’s (1985, 2000, 2007 lexicalization pattern approach. Examples (1) through (6) show that the site component of a stationary location event is always expressed by adpositions. Besides, the semantic component of locatedness conflated with a position of the figure can be encoded by a positional locative verb. This suggests that Amharic uses a satellite-framed pattern in static locative constructions. However, I do not claim that Amharic is a satellite-framed language because the material discussed in this paper does not include expressions of movement that are subsumed under Motion event. Regarding the variability of patterns of Motion constructions between static and movement expressions, the present study shows that Amharic uses satellite framing for static location. For movement expressions, other studies such as Wakasa (2016) have argued that Amharic uses a verb-framed strategy.

4.1 Spatial prepositions

In the literature, the prepositions of Amharic including the spatial prepositions are treated in two different ways: (1) as independent words (Anbessa and Hudson 2007: 46–47; Baye 2016: 93; Leslau 1995: 597–616) (2) as semantic case markers (Baker and Kramer 2014: 144; Hetzron 1970: 308–315; Tremblay and Kabbaj 1990: 175). In this study, I treat them as independent words classified as prepositions.

In Amharic, basic location can be expressed by using the spatial prepositions kə, tə- and ɨ- ‘at’.^[9] These prepositions have a general locative function; they do not show specific types of topological relations between the figure and the ground. This means that they convey a dimensionless position of the figure. The distinction among the three involves dialect variation. Speakers of the Godʒdʒam variety (GA) in rural areas usually use tə- or ɨ- interchangeably in situations where the noun denoting the ground begins with a consonant. In such a case, they serve as free variants. When the ground noun begins with a vowel, they may use tə-. Those speakers who live in towns and have exposure to ‘standard’ Amharic (henceforth SA) tend to use the preposition kə-. For illustration, consider utterances (7) and (8).

As shown in example (7) above, the ground noun t’ərənbeza ‘table’ begins with the consonant /t’/. In such a case, those speakers who do not have exposure to SA use the locative prepositions tə- and ɨ- interchangeably as in (7a). In example (8) above, the ground noun angət ‘neck’ begins with a vowel and speakers do not use the morpheme ɨ-, instead they use tə-. On the other hand, speakers who have exposure to SA use kə- both in consonant- and vowel-initial words, as in (7b) and (8b).

There is also another locative preposition bə- ‘at’, which has limited contexts of use. Based on the data I obtained, it is marginalized in BLC. It can be deployed in situations when a figure is firmly attached to the ground (e.g., stamp on the envelope) as well as when a figure is physically separated from the ground (e.g., fence around the house). However, such topological relations are not good candidates for BLC treatment (Levinson and Wilkins 2006a: 515). In such cases, the locative expression involves resultative constructions, as in (9).

In example (9), the verb tələt’t’əf- ‘be glued’ is a passive resultative verb expressing the result of gluing. In such an instance, the preposition bə- ‘at’ shows that the noun bost’a ‘envelope’ has the function of location at which tembɨr ‘stamp’ is placed.

In a copular construction, bə- ‘at’ can be used in contexts in which the ground is a bounded container and emphasis is on the container.

In this example, the preposition bə- shows that the ground bɨtʃəre ‘glass’ serves as a container of t’əlla ‘local beer’. In such a case, speakers emphasize the container. The container itself is the location of t’əlla. However, bə- does not frequently occur in typical locative expressions in the Amharic variety of Godʒdʒam.

4.2 Spatial postpositions

Amharic uses a considerable number of spatial postpositions to express the part of the ground where the figure is located. ^[10]The postpositions may encode specific relational positions with reference to the ground. Amharic spatial postpositions may encode a region or a part of the ground at which figure is located (see Table 3).

4.3 Combinations of spatial adpositions in basic locative construction

As discussed above, Amharic uses spatial adpositions (prepositions and postpositions) in BLC. The prepositions designate a general location of a figure while the postpositions can render the figure’s specific location, spatial proximity, and region with respect to the ground. These adpositions may also be used in combination (cf. Anbessa and Hudson 2007: 80–81). Since locative prepositions do not specify the location of the figure with respect to the ground, they co-occur with spatial postpositions. For example, a scene involving ‘cup on the table’ (picture 01 of TRPS) can be described as follows.

The basic location in (21) is constructed by using the general locative prepositions tə- or ɨ- with the postposition laj. The prepositions tə- and ɨ- express a spatial relation while the postposition laj denotes the spatial region of the ground object where the figure is located. The prepositions can be omitted without bringing about any appreciable semantic difference (compare example (21) with (22)).

Similar to the prepositions, postposition such as laj may be omitted in situations when the ground is a familiar location (e.g., home, market, church, school). In this case, site or the particular region of the ground is not specified as in (23b) or it can be interpreted from the context (from the nature of the ground) as illustrated in (23c), examples from introspection of author’s native speaker intuition. In such cases, only a general imprecise location of the figure vis - à - vis the ground is expressed.

The prepositions tə- or ɨ- may be used without the spatial postposition laj as in (23b). In example (23c), there is no preposition or postposition to refer to the ground. However, (23c) is grammatical and the noun bet ‘house’ is interpreted as a location, a kind of elliptic construction that is perhaps made possible because the response is framed by the interrogative pronoun jət ‘where’ in (23a).