Academic journal article Studies in African Linguistics

Locationals in Oromo1

Academic journal article Studies in African Linguistics

Locationals in Oromo1

Article excerpt

This is a study of the locational structures of Oromo. A range of syntactic constructions types is considered within a single synchronic grammaticalization schema. Speaker choices of particular structures within discourse are also identified and explored. The primary data are drawn from the Guji dialect, with reference to data from other dialects that are attested in the literature. Most of the morphological marking that is found across these locationals is consistent in all Oromo speech communities, and, although there is some variation in some particular lexemes across the dialects, the inventories of locational lexemes are interlocking and nearly entirely overlapping.

In Oromo, a Lowland East Cushitic language of the Afro-asiatic family, there is a range of constructions that speakers use to locate things. This study explores the nature of these constructions, their relationships within Oromo grammar, grammaticalization principles that relate to them, and the correlation between the form of locational structures and the status of referents and information in discourse. An examination of locational structures in the Oromo dialects reveals grammaticalization within the synchronic system. The data illustrate how a speaker?s choice of locational structures is driven by pragmatic considerations. Examining locational structures from the perspective of discourse establishes insights into the nature of Oromo grammar that are obscured if these constructions are only considered at a syntactic level. The data provide evidence that the degree to which information is shared between interlocutors within a discourse determines appropriate choices that allow for coherent interpretations of locational messages and show how pragmatic status functions to create grammaticality.

The terminology used for locational structures primarily follows that suggested by Talmy (1975:181-2).2 The figure refers to the object that is located or is moving, and the ground the object or place with respect to which the figure is located or moving. The figure and ground are typically nominals. The site identifies the location of the figure with respect to the ground, and the path specifies the course through which the figure is moving in relationship to the ground.3 In Oromo, the site or path may be expressed by a postposition, a relational noun in a genitival construction, an affix, a directional, a preposition or some combination of these.

Five major regional dialects of Oromo in Ethiopia and three in Kenya are located in Figure 1. Boraana is the dialect of southeastern Ethiopia and northeastern Kenya. Guji is spoken in the region of Ethiopia northwest of the Boraana. Harar (known also as Ittu) is the dialect of the most northeastern Oromo area of Ethiopia. Tulema (also identified as Shoa) is spoken in the central area surrounding Finfinnee (renamed Addis Ababa "new flower? when the Amhara seized control of the area). Wollegga (sometimes called Macha) is the westernmost variety. Gabra, Orma, and Waata are spoken in Kenya.

In this study examples of the various types of locational structures are provided, common formal markings found on the different types of constructions are identified, and grammaticalization principles that apply to specific examples and structures are discussed. Finally, the effect of the status of information in the discourse on the choice and form of locational structures is examined. The primary data are from the Guji dialect of Oromo. The data are drawn from the Lowland Guji narrative in Yaachis and Clamons (2009), which is a transcript of a videotaped "near death? account, and also from data constructed by Mi?eessaa Yaachis, a Lowland Guji speaker, on the basis of native intuitions. Mekonnen Abakore, a Highland Guji, has verified the data from the narrative and the constructed conversation data. Although he is Highland Guji, rather than a Lowland Guji, no differences were found in his intuitions about the locational constructions considered here.

Information about locational structures in the other dialects of Oromo is included when it is attested in the literature. Stroomer (1995) provides descriptions of Boraana Oromo, Ali and Zaborski (1990) and Owens (1982, 1985) describe Harar Oromo, Moreno (1939) and Bender, Eteffa and Stinson (1976) report on Tulema Oromo, and Gragg (1976, 1982) and Griefenow- Mewis and Bitima (1994) and Griefenow-Mewis (2001) portray Wolleggan Oromo. Stroomer (1987) documents Gabra, Orma and Waata Oromo of Kenya.

There are five syntactic locational construction types that function to locate objects in Oromo: postpositions, relational nouns in genitival constructions, affixes, directionals, and the preposition gara. Several of the construction types in Oromo illustrate the grammaticalization cline introduced by Lehmann (1995(1982): 25), which he describes as ""a theoretical construct along which functionally similar [sign] types are ordered according to their degree of grammaticality as measured by certain parameters . . ." and suggests the ordering:

The "positions ? on this cline are not discrete, but represent conventional categories across which the grammatical structures may be ordered, based on certain paradigmatic and syntagmatic parameters. The constructions on the left are most lexical, least grammaticalized; those on the right are most grammaticalized, least lexical. Lehmann?s grammaticalization cline and the parameters suggested in Lehmann (1985) and the principles in Hopper (1991) illuminate how certain Oromo locational constructions can be evaluated as more or less grammaticalized and how the relationships that pertain between various construction types may be assessed. For instance, one of the principles introduced in Hopper (1991) is layering. He proposes that "[w]ithin a broad functional domain, new layers are continually emerging. As this happens, the older layers are not necessarily discarded, but may remain to coexist with and interact with the newer layers." (1991:22) Scholars working on grammaticalization phenomena often consider language diachronically. Although some lexical origins are noted, this study focuses on the synchronic structure and the grammatical relationships between different construction types across the cline. The data from Oromo, with its five locational construction types, provides illustrations of layering as it is emerging in the current Oromo grammar, without reference to historical processes.

1.1 Postpositions. In Guji, as well as in all other attested dialects of Oromo, there is a set of postpositions that mark the site in locational structures. The members of this set of locationals always follow the nominal construction that instantiates the figure. In the example in (1), the postposition jala "under? is the site, and ifi "myself? is the ground where the figure, ani "I?, did not look. 4

(1) ani waan ifi jala hin laalatinii-

I+ni since myself under not look

"Since I didn?t look under myself"?

fse"lf "5

In the next example in (2), the postposition keessa "in? provides the site and loon "cattle? names the ground where the figure, nuuti" "we"?, is located.

(2) nuuti " ani ijoollee obboleeyyan tiyyaa woliin

we " I+ni children siblings my together

loon keessa kama tissiisatti jirru"

cattle in while pasturing were

"".while the children, me and my brothers and sisters, were pasturing the cattle"?(11)

The postpositions of Guji are listed in Table 1.

The postpositions listed here are also attested in the other dialects with the following exceptions: duudii and faana are not attested as postpositions in dialects other than Guji, although faana occurs as "track? or footprint? in Harar and Boraana; gula and haga are not attested in Wollegga or Tulema; jala is not attested in Orma and Waata. The compound fuuldura is also found in Wollegga, but is not attested in the other dialects, but dura and fuula "front? are. In Boraana, duduuba is the lexeme for "behind?. Not found in Guji, but attested in Harar, Tulema, and Wollegga is booda "after?, in Harar and in Wollegga is teella "after, behind?, and in Wollegga is duukaa "after?. The reciprocal form woliin "together? is also found in Guji and Boraana as a postposition, "between? and is attested in the Harar and Wollegga as wajjiin. Bukkee is attested only in Wollegga and Guji. Although historical work on the dialects has yet to be done, these few variations found in the inventory of postpositions across the different groups of the Oromo speech community may reflect emerging grammaticality as contact with other speech communities and creative language activity shape the grammar of the language uniquely in the various regional communities.

1.1.1 Path Marking on Postpositions. Postpositions that indicate movement along a path are marked with a long final vowel. For example, jala-a "from under?, that represents the path of "escape from? in the example in (3), has a long final a, while in example (1), jala "under?, that represents the site, has only a short vowel.

(3) Tanaaf" makaraa adda addaa waan hedduu jala-a baana.

therefore hardship front front since many under PATH escape

"Therefore".. we escape". many different kinds of hardship.?

Similarly, keessa-a "into ? in (4), representing the path through which "I ? dropped, is marked with a final long vowel, while keessa, denoting the site in example (2) has a short final vowel.

(4) Tanaaf" utaalee obbaa keessa-a bu?e.

therefore jump reeds into.PATH dropped

"Therefore... I dropped into the reeds.?

The more metaphorical path by which sight cannot be gained in example (5) is also marked keessa-a with a long vowel.

(5) Obbaan kun daggala marraa gudda?aa,

reeds this bulrushes grass high

nami keessa-a hin mudhdhatu.

person+ni in PATH not be.seen

"These reeds, the bulrushes are so high you can?t be seen.? (18)

Griefenow-Mewis and Bitima (1994:128-129) analyze this lengthening on locationals as "von, aus, or heraus (from, out, or out of)?, but our data suggests a less specific, more contextualized path. Owens (1985:116-17, 127) analyzes this lengthening as dative case form. Stroomer (1995:99) describes it as a "linker clitic? denoting ablative. Although lengthening of a final vowel is found across multiple construction types with a number of distinct functions in Oromo, as analyzed extensively in Stroomer (1995:94-118), what is of note here is that it indicates motion across a contextually determined path type in all of the locational construction types, across the cline.

1.2 Relational Nouns in Genitival Constructions. Relational nouns in genitival constructions also function to identify site in Guji. This is a common strategy across languages. Heine and Reh (1984:101) find that "[m]ost, if not all, African languages use the transfer strategy to express prepositional concepts by means of genitive constructions." In the example from the Guji given in (6), the figure, loon "cattle? "edge?, of the ground, laga-a "valley, river, stream, water?.

can be sought at the site, afaafa

(6) A: loon eessa jiran?

cows where are

"Where are the cows? ?

B: afaafa laga-a barbaadi!

Edge water.GEN search

"Search by the edge of the water! ?

In example (7), " he?, understood from the verb, is the figure, horaa "salt lake ? is the ground and karaa "way, road? is the site.

(7) yo worra galu, karaa hora-a gale.

when group returning way salt-lick.lake.GEN returned

"When he was returning home, he returned by way of the salt-lick lake.?

1.2.1 Path Marking on Relational Nouns in Genitival Constructions. The final long vowel to indicate path, as on postpositions, as well as on locational affixes and directionals discussed below, is vacuous in genitival constructions with relational nouns because the final vowel in the genitive form is already long. Although path is marked, because the genitive marking is coexistant with path marking, the actual manifestation of both equates with that of either.

1.3 Locational Affixes. Two affixes, -tti "at, in? and -rra "on, above, over?, also serve to identify sites in locational structures. In the following example, -tti "in? functions as the site where the figure, ulee "stick?, was in the ground, harka "hand?.

(8) ka ulee dheertuu takka harka- tti qabatee

who stick long INDEF hand in had

""who had a long stick in his hand"?


In example (9), -rra functions as the site where the figure, simbirreen "birds?, circle over the ground, mana "house?.

(9) simbirree- n mana- rra marti

Birds SUB TOP house above circle

"Birds circle above the house.?

In Guji, these locationals always occur as affixes, although speakers can identify them as representative of the full forms, itti and irra, and will cite these as full forms in isolation. In other dialects of Oromo the locationals may appear as full forms. Gragg (1982:227) identifies the full form as a postposition, but indicates that it is usually enclitic. Owens (1985:115) claims that the affix has been derived from the postposition in Harar. In Guji, as well as other Oromo dialects, there are verb particle constructions created with the full forms itti and irra, as in (10)-(13). Heine and Reh (1984:135-43) point out that verbal derivative extensions are grammaticalized from adpositions and that derivatives such as these in Oromo tend to be on the opposite end from inflections, as they are here.

(10) itti qaba.

at/in have

"I hit it.? (Lit. "I have at (it).?)

(11) itti bara.

at/in learn

"I practice.? (Lit. "I learn at (it).?)

(12) irra jira.

on/above is

"It is bigger.?, "It is newer.? (Lit. "It is on/above (it).?)

(13) irra-a chab-e.

off.PATH broke 1.S

"I broke away from it.? (Lit. "I broke off from (it).?)

According to Lehmann (1985:306-7), the more grammaticalized a sign is, the less autonomous it is. The less phonological size a sign has, the more bonded it is, the more obligatory it is, and the less able it is to fill a different slot in a construction, i.e., the less autonomous it is. These Oromo data reflect two of the principles of grammaticalization introduced in Hopper (1991): divergence and persistence. By these criteria, -tti and -rra are more grammaticalized than the postpositions. According to Hopper (1991:22), divergence can be seen when a lexical form is grammaticalized as a clitic or affix and ""the original lexical form [remains] as an autonomous element." The coexistence of the Guji the locational affixes -tti and also -rra and also the full forms itti and irra, as shown in the examples above, illustrate this principle of divergence. This reflects the more grammaticalized affix forms as they are related to the full lexical forms which are nevertheless still present in the grammar. Hopper (1991:22) indicates that persistence is demonstrated in a form that is more grammaticalized than a corresponding lexical form when ""some traces of its original lexical meanings " adhere to it." The Guji locational affixes -tti and -rra carry meanings related to those of the full forms itti and irra, as shown in the examples above, and thus illustrate this principle of persistence. Although the affixes are more grammaticalized, both the reduced and full forms share lexical semantic values.

1.3.1 Path Marking on Locational Affixes. As is the case with postpositions, the locational affix -rra may also be marked as path by the length of the final vowel, as illustrated in example (14), as opposed to the example in (9).

(14) " ifi -rra-a buqqifatiisa hin dandeennee

"myself out.from PATH pull not could

"" I could not" pull (the thorns) out of myself"?

We have found no instances of lengthening on the -tti affix, making it a true exception.

1.4 Directionals. The directionals gadi "down? and ol "up? also locate the site in locational constructions. In the following exchange in (15), the "donkey? is located "down the mountain?.

(15) A: Haree-n eessa jirti?

where is

"Where ? s the donkey? ?

B: Gaaraa- n6 gadi jirti.

mountain n down is

"It?s down the mountain. ?

In (16), ol and gadi establish the location where the figure, harree "donkey? is found ascending and descending on the ground, gaara "mountain?.

(16) Haree-n gaara ol baate, gara kaaniin gadi buut-e.

donkey SUB TOP mountain up ascend toward that.other down descended

"The donkey plodded up the mountain and went down the other side.?

Owens (1985:60) and Gragg (1982:159) identify gadi and oli/ol as adverbs in Harar and Wollega. Griefenow-Mewis and Bitima (1994:95-96) and Griefenow-Mewis (2001:52) treat them as postpositions in Wollega. Although they may both occur in conjunction directly before the verb, we consider them separately from postpositions, because they are otherwise

inseparable from the verb. In examples (17) and (18), the impossibility of moving the adverb between ol and the verb is illustrated.

(17) Haree- n suutaan gaara ol baate.

donkey SUB.TOP slowly mountain up ascend

"The donkey plodded slowly up the mountain?

(18) *Haree-n gaara ol suutaan baate.

A postposition may, on the other hand, be separated from the verb, as illustrated in (19).

(19) Haree- n moonaa keessa suutaan seente.

donkey SUB.TOP kraal in slowly entered

"The donkey slowly entered the kraal.?

As with itti and irra , directionals operate as verb particles, with particular semantic values when in combination with certain verbs across dialects, as illustrated in examples (20)-(23).

(20) Inni gadi fulla?e.

he down appeared

"He appeared all of a sudden.? (Lit. "He appeared down.?)

(21) Gadi chadiisi!

down quiet

"Shut up!? or "Settle down!? (Lit. "Quiet down!?)

(22) Ol te?a.

up be.1S

"I am more than.? (Lit. "I am up.?)

(23) Ol qaba.

up have.1.S

"I glorify/exaggerate something.? (Lit. "I have up.?)

As with the locational affixes, these particles reflect persistence, since the directional displays a semantic value very similar to that of the particles in these particle verb constructions.

1.4.1 Path Marking on Directionals. As with postpositions and locational affixes, gadi may be marked with a long vowel that identifies motion along a path, rather than a static locational site, as illustrated in (24), in contrast with the example in (15).

(24) Inni gadi-i fulla?e.

he from.below PATH appeared

"He appeared from down below.?

1.5 Preposition gara. Guji, like all other Oromo dialects, has the preposition gara "to, toward?.7 The following examples illustrate the use of gara.

(25) gara mana barumsa-a deeme

toward house teaching PATH came

"He came to the school.?

(26) Achii-n duuba fuudhanii gara hori-?ii

there+n behind they.took to cattle PATH

"Afterwards they took and returned me to the cows"?

na deebisanii"


(27) "isaan biraa utaalee dhokatiisaaf gara laga bisaan-ii,

Them from jumped toward gully water PATH

"I jumped away from them to hide, [toward a gully of water]"?

This preposition is also used to locate an approximate point in time, as illustrated in (28).

(28) Mani barumsaa sun"gara saatii afur-ii fudhata.

house+ni teaching that about hours four PATH take

"That school is about four hours away... ?

It is extremely rare for a verb-final language such as Oromo with its typologically expected postpositions to also have prepositions, even a very few prepositions.8 Owens (1982:60-62, 1985:128) analyzes gara as a noun, not a preposition, and Griefenow-Mewis (2001: 51) proposes that it be treated as a noun. Bender, Eteffa, and Stinson (1976:139) do suggest that gara is probably derived from a noun and Gragg (1976:184) claims that prepositions in Oromo are derived from nominals. But gara only appears in locational constructions, never as a noun in subject or object phrases in Guji. It is identified a "place, side, direction? in Boraana (Stroomer 1995:178) and in Kenyan Boraana, Orma, and Waata (Stroomer 1987: 311). It appears to be chosen on the basis of pragmatic status in Waata, as discussed in section 3.2 below.

Movement is an inherent semantic value of all constuctions with gara . Prepositional phrases with gara therefore always have a long final vowel. This suggests that path marking is phrase final, rather than lexeme final in all constructions. This phrase final marking is consistent with other suffixes in Oromo, as identified in Owens (1980:155 ff) for oblique case suffixes and also in Clamons, Mulkern, and Sanders (1991, 1993) for the case marking on non-topical subjects. Owens (1985:128) argues that Harar Oromo has no prepositions, but that those lexemes that look like prepositions are nominals in a genitival construction. He analyzes the phrase final marking as genitive, while it is analyzed as path marking here.

2 Combinations of Locationals

In Oromo, as is common in other languages, locationals are frequently used in combination, thus providing a more complex description of path and site. Postpositions and affixes are combined as in the examples in (29) and (30).

(29) "nulle eegee duuba?rra miilumaan baaddiyaa keessa worra-tti galle.

"we.also tail behind [from] by.foot country through family to left

"finally we also left the school and went behind on foot through the countryside to our

family.? (9)

(30) "jia lamaa fi sadi keessa-tti dhufanii achitti na ilaalan.

"moon two and three come there me see3PL

""they came to see me there every two or three months.?

Without the affix, the meaning changes slightly, as in (31).

(31) "jia lamaa fi sadi keessa dhufanii achitti na ilaalan.

"moon two and three in come there me see3PL

""they came to see me there within two or three months.?

Gragg (1976: 184) and Griefenow-Mewis (2001: 51) analyze these as (parapositions).

A directional and an affix may also be combined, as in (32).

(32) Inni muka- rra-a gadi bu?e.

he tree from PATH down dropped

"He dropped down from the tree.?

Multiple directionals can be used to indicate path and orientation as in (33).

(33) Inni gadi-i ol yaabe.

he down PATH up climbed

"He climbed down from above.?

The preposition gara occurs with postpositions, as in (34), an example from Yaachis (n.d.: Mark 1:19), as well as with directionals, as in (35).

(34) ".akkuma tiyyoo gara dura -a demeen"

just as little towards in front PATH having gone

""having gone a little further thence".?

(35) "gara laga kaanii gadi caafamee

toward gully that down turn

""I turned down towards that gully"?

The preposition may also co-occur with a locational affix, as in (36) and (37). Gragg (1976:184) analyzes this combination as a paraposition in Wollegga Oromo, while Griefenow-Mewis and Bitima (1994:96) analyze gara"-tti "nach (to, the direction to)? as a circumfix.

(36) "isaan baqatanii gara badda?aa gosa dhibii- rra-a

they fled to forest area other from PATH

fagaatanii qubatanii jiran"

far.away settled were

"" they had fled to the highland area far away from the others and settled"? (9)

(37) "gara teessoo teennaa-tti"

in area

""in that village of ours"?

2.1. Locationals in idiomatic expressions. Locationals are also found in combination with other elements in idiomatic expressions, as with irra and the quantifier in example in (38) and the reduplicated preposition gara in (39).

(38) "yennaa lolaa tana keessa irra caalaa bidhdhaan...

during war this in above all suspicion

""during this war, very often there was a different kind of suspicion"

(39) "lola gara garaa, bineensa gara garaa".

war different.kind wild animals different.kind

""different kinds of war and different kinds of wild animals"

This occurrence of these in frozen reflexes again reflects the principle of layering, with both layers coexisting in the grammar. Gragg (1982:167) provides examples for Wollegan Oromo with reduplicates of gara as in the examples in (40) and (41).

(40) garaagara as adverbial "apart?

(41) garaagara ta'a as verbal "be separated?

This illustrates the principle of persistance, with the semantic value of movement, in these cases "away from,? incorporated in the semantics of the idiom.

3 Discourse and the Forms of Locationals

Lehmann (1985:314) asserts that ""the linguistic system [is]"created by language activity." Oromo data from Guji illustrate how the status of information in a discourse determines the form, ordering, and privilege of occurrence of elements in locational structures. Locationals are eligible for the phonopragmatic marker n(V) that appears phrase- finally on constructions that signal that referents or information are accessible to interlocutors within the domain of a discourse. Also in Guji Oromo, relational nouns in genitival constructions that mark a site or path that can be identified by both speaker and addressee are postnominal, and in this position, they have the grammatical status of postpositions.

3.1 Pragmatic n marking on Locationals. Yaachis and Clamons (2009) identify a phonopragmatic marker, n(V), that is found finally across a range of morphological, syntactic, and semantic expressions and reliably signals accessibility of referents or information. Eligible elements of subject noun phrases that establish referents that are at least familiar to the speaker and addressee, pronominal forms for specific referents, spatiotemporal forms that index times and locations that can be identified by interlocutors within a domain of discourse, complementizers that signify recoverable logical relations, and topical arguments are consistently marked as n(V) on the right edge of expressions. The n(V) marking on these constructions, with some formal variation, is found across the dialects of Oromo. Example (1), repeated here as (42), exhibits the n(V) on ani, which is a topic subject, uniquely identifiable to speaker and audience and on waan, which follows the accessible information in the sentence.

(42) ani waan ifi jala hin laalatinii- f "

I+ni since myself under not look self "

"Since I didn?t look under myself"?

In Guji Oromo, this pragmeme is also found on the right edge of all types of locational constructions and indexes an anaphor or an identifiable location. Introduced in Yaachis and Clamons (2009) is the notion of a pragmeme as a most simple phonic sign found systematically across syntactic and morphological structures to reliably indicate pragmatic status of referents and information status. This differs from the notion of pragmeme as a matrix of features that Mey (2001:222) uses to characterize both the activity and context of speech. The marking of a locationals with this n pragmeme is illustrated in the following examples. The example in (43) represents the meaning of bira "by, near, with? as the site of the ground isa "him?, which locates the speaker "at his home? for the addressee.

(43) isa bira jira.

him with be 1S

"I am at his home.? (Lit. "I am with him.?)

In the following examples in (44) and (45), there is an n marker on the pronouns that represent the ground. Yaachis and Clamons (2009:16:(28)-(29)) point out that this is appropriate for ""referents that are in focus, that is, that are in the central awareness of the interlocutors"" as in these two examples.

(44) isaa-n bira jira.

him TOP with be 1S

"I am with him.?

(45) isaan-iin bira jira.

them TOP with be 1S

"I am with them.?

Yaachis and Clamons (2009:172) also point out that "" the instrumental object form for an in-focus referent may be an anaphoric -n that appears as a clitic on the right edge of the adposition"?, as it appears on the postposition bira in example (46).

(46) isa biraa- n jira.

him with it be 1S

"I am with him with it.?

"I am with him doing it.?

Hopper and Traugott (1993:2) maintain that ""grammaticalization is primarily a syntactic, discourse pragmatic phenomenon, to be studied from the point of view of fluid patterns of language use." This is illustrated in this example. The n pragmeme on referring expressions, spatiotemporal forms, and complementizers across the dialects of Oromo, and on locational constructions in Guji, signals that some referent or information is shared by the interlocutors because it is in the immediate domain, has been mentioned in the domain of the discourse, or is inferable from the immediate or discourse context. Example (46) may carry either the meaning "I am with him with it.? or the meaning "I am with him doing it.? The difference in the readings depends on what information is shared by the interlocutors at this point in the discourse. In the first case, the addressee is assumed to be able to retrieve into consciousness what it is that "I have with me?; in the second, the addressee is assumed to know "what I am doing?.

This is further demonstrated in the following examples. In (47), the n marker indicates an accessible site. In (48), the n on the postposition indicates the accessible site, "up to them?, and the n marker on the verb signals an object that can be uniquely identified by both speaker and addressee.

(47) Isaan biraa- n baye.

them near get to

"I caught up to them.?

(48) Isaan


biraa- n


bayee- n. get to obj

"I caught up to them with it.?

Example (49) also illustrates the use of multiple n?s to signal both locational and referential information that can be retrieved by the interlocutors.

(49) Daa?ima fayyaa hin qanne baadhee isaan biraa- n bayee -n.

child well not have carrying them near caught1S

"I was carrying a sick child and caught up to them with it.?

Directionals may also be marked with n to signal the pragmatic status of referents and information. The following exchanges demonstrate how knowledge that is either elicited from an addressee by the speaker, or contrariwise is assumed to be already shared at this point in the discourse, conditions the presence or absence of this phonopragmatic marking on responses. In the first exchange, in (50) speaker A implies only the knowledge of a referent "he?, that is shared with addressee B, has "veered off?, and requests information about the location. Therefore in the response, the site, natta gadi "down towards me?, has no pragmatic marking.

(50) A: Eessa gore?

where veered 3S

"Where did he veer off? ?

B: Inni karaa san dhufee natta gadi gore.

he road that came down veered 3S

"He veered that way and veered down towards me. ?

In this next exchange, in (51), A seeks more information about where exactly "he? veered off, but includes the explicit anaphoric object marker, n, on the question word, signifying that both interlocutors have knowledge that "he? has "it?, whatever it may be. B therefore marks natta, the site, with n, anaphoric for "it?.

(51) A: Eessaa-n gadi gore?

where it down veered

"Where did he veer off with it? ?

B: Inni karaa san dhufee nattaa-n gadi gore.

he road that came OBJ down veer

"He came down the road towards me with it. ?

In the following exchange, in (52), the anaphoric n for the object that is assumed by speaker A to be identifiable by the addressee is marked on gadi in the question. Thus the anaphoric object is marked on the directional in B?s answer.

(52) A: Eessa gadii-n gore?

where down it veered

"Where did he turn off with it? ?

B: Inni karaa san dhufee natta gadii-n gore.

he road that came down it veered3S

"He veered off down that road towards me with it. ?

In this final exchange, in (53), speaker A signals, by the n on essa, the awareness that "he? veered off with "it?, but is requesting more specific information about the site. B provides this with natta gadii-n, and includes the implicitly shared information that "he? has "it? with him, a referent identifiable by both A and B, now marked on the verb.

(53) A: Inni essaa-n gore?

he where to veered3S

"Where did he veer off to with it? ?

B: Inni karaa san dhufee natta gadii-n gore- en.

he road that came down to veered3S OBJ

"He veered down that road towards me with it. ?

The final n is also found on the temporal constructions in (54) and (55). In both of these examples the narrator has also established time frames against which these temporal sites can be located and signals this to the audience with the final n pragmeme.

(54) Ani ammoo gara bulii sadii-afurii -tii -n dura-tti,

I+ni but before night three-four.PATH [SEP] n

"But I had, just three or four nights before, "?

(55) "gara saati lamaa -tii -n duubatti

for hour two.after PATH [SEP] n said

"" about two hours after "I said?

Griefenow-Mewis and Bitima (1994:96) analyze gara"-tii-n as "von"her, aus der richtung (from hence, from out of the direction)? as a circumfix in Wollegga Oromo, where the interlocutors can mutually identify the source location from which the movement is initiated. Although there is no attested data showing the final n on locational constructions with this anaphoric reading in other dialects of Oromo, it is found anaphorically on the verb in Harar, illustrated in the data in (56) from a story.9

(56) Intal-aa magaalaa dhufe. Tokko arka -n.

girl SUB market PATH came one M sees OBJ

"The girl leaves the market. A man sees her.?

The examples in this section illustrate how the phonopragmatic n marker is used by interlocutors with locational structures in Guji to signal what the status of a referent or site is assumed to be, just as it is used across dialects in the other construction types that are discussed in Yaachis and Clamons (2009). In various contexts, the n signals different semantic and grammatical values, but consistently identifies information that is assumed to be accessible to the interlocutors at the point in the discourse that n is selected. Yaachis and Clamons (2009:17) point out that this marker is often a portmanteau form, carrying a complex of morphological, syntactic, or semantic information in addition to the indexical meaning it signifies in the domain of discourse. In (42), for example, the ani form of the first person personal pronoun signals both the grammatical subject and the uniquely identifiable status of the referent. In (48) the n carries the grammatical value of instrumental as well as an anaphoric value. Although this pragmeme carries different information, depending on the grammatical context, it always identifies referents or information that a speaker assumes can be retrieved by the addressee at the point of the discourse where it is used.

3.2 Relational Nouns as Postpositions. As pointed out in section 1.2 above, it is common in African languages for locations to be designated with noun noun genitival constructions. However, in Oromo a subset of relational nouns that signify sites in genitival constructions also are found as postpositions when the site is accessible for all interlocutors at that point in the discourse. The discussion provided below illustrates the genitival construction with the site, afaafa "edge?, and the ground, laga "valley, river, stream, water?. In the following example, (57) that repeats example (6) in section 1.2 above, afaafa "edge? is a relational noun in a genitival construction with laga-a "of the water?.

(57) A: loon eessa jiran?

cattle where are

"Where are the cows? ?

B: afaafa laga-a barbaadi!

edge water GEN search

"Search by the edge of the water. ?

Hopper and Traugott (1993:67) identify communicative negotiations as one of the factors that drive grammaticalization, and Traugott (2005:634) argues that pragmatic strengthening arises from the cognitive and communicative realities of speaker-hearer interactions. In example (58), a second exchange between A and B illustrates how the more grammaticalized postposition for the site afaafa "edge? is chosen once the ground, laga "valley, river, stream, water? is established for both interlocutors in the domain of the discourse.

(58) A1: loon harkaa na badan.

cows hand me lost

"I lost the cows. ?

B1: laga kana keessa barbaadi!

valley this in search

" Search in this valley! ?

A2: laga kana eessa middee?

valley this where exactly

"Where exactly in the valley? ?

B2: laga afaafa barbaadi!

water near search

"Search near the water!?

In the first exchange, in (57), speaker B cannot assume that speaker A knows to look in the "valley?, and appropriately chooses the genitival construction. In the second exchange, in (58), the "valley, river, stream, water? has already been established as the site for both interlocutors and the use of afaafa as a postposition is thus appropriate.

Similarly, in the first exchange below in (59), jidduu "middle? is used in a gentival construction to indicate the site where the dog can be found.

(59) A: saree-n eessaa jira?

dog SUB TOP where is

"Where is the dog? ?

B: jidduu mana-a jira.

middle houses GEN is

"It is in the middle of the houses. ?

In the following exchange in (60), however, where the site and ground are both introduced in the question, and therefore are immediately identified by A for both speaker and addressee at this point in the discourse, jidduu "between? occurs as a postposition.

(60) A: maanti mana jidduu jira?

what houses between is

"What is between the houses? ?

B: sarree-n mana jidduu jira

dog SUB TOP houses between is

"The dog is between the houses. ?

In example (61), wodhakkaa "in the center of, in the middle of? precedes Oromo 'ootii "of the Oromo?, Guji'iitii "of the Guji?, and Somale'ee "of the Somali? in noun noun genitival constructions, but is postpositional "between? following the clarification of Digoodi'ii and Maryana as "people who are called Digoodi or Maryaana?, referents that have thus been made identifiable for the audience at this point in the narrative.

and or

Maryaana worra

Maryaana people

(61) Gizee hedduu ammoo wodhakkaa Oromo?ootii fi yokiin

Time AMH many but between Oromo

Guji?iitii fi wodhakkaa Somale?ee yokiin Digoodi?ii,

Guji and between Somali or Digoodi

jedhanu kana wodhakkaa lolá -tti ka?a.

called this between fight SUBJ EMPH get.up

"Often between Oromo and or Guji and Somali or those people who are called

Digoodi or Maryaana, a war breaks out.?

The relational nouns of Guji that can also occur as postpositions are given in Table 2.

Critically, although many other relational nouns can occur in genitival constructions, only certain nouns may also occur as postpositions. For example karaa "way, road? in example (7) in section 1.2, cannot occur as a postposition. Eligible relational nouns in Guji Oromo that identify the site in the noun noun genitive constructions are homophonous with the corresponding postpositions and carry a similar conventionalized semantic value. However, marking on the genitival nouns indicates only "genitive?, while the postpositions may be marked with lengthening that indicates path when they occur with verbs of movement, and they may also bear the n marking characteristic of expressions with familiar discourse status, as discussed above in 3.1. This difference is consistent with Hopper?s notion of de- categorialization, whereby more grammaticalized forms don?t have marking of categories such as noun and but do exhibit those of "secondary categories? such as adpositions (Hopper 1991:22, 30-31). Furthermore, as nominal forms they are less restricted syntactically, occurring in other nominal constructions, with modifiers, etc. As nouns they are more lexical, as postpositions, more grammaticalized. This critical syntactic difference is illustrated in the following exchange in (62).

(62) A1: bulluk koon tiyya meeti ?

white blanket my where

"Where is my bullukkoo? ?

B1: afaan kana ilaali!

mouth this look

"Look for it around here! ?

A2: afaan kana eessa middee?

mouth this where exactly

"Where exactly? ?

B2: boolla qaayyaa afaan (*kana)

incensing hollow mouth * this

"By the incensing hollow. ?

In the response in B1, the noun afaan "mouth? can be modified with the demonstrative. In B2, where afaan now is accessible and functions as a postposition, it cannot. Again, as with afaafa, this lexeme is only appropriate as a postposition in the discourse once both interlocutors share knowledge of the location already indicated in B1. As Traugott (2005:631) points out, ""grammatical phenomena that "serve interface functions with discourse? must be in the grammar because they entail structures that occupy syntactic positions." The choice between a noun-noun construction and a noun postposition construction, and the selection of the n pragmeme, are dictated by the assumed status of the encoded information for the interlocutors. This provides evidence that the pragmatic status of information at a given point in discourse determines the grammatically appropriate form. Importantly, while the set of relational nouns that may identify location in genitival constructions is open-ended, only a restricted set of these nouns also occur as postpositions.

The examples here are from the Guji dialect, but similar patterns are evidenced in other varieties of Oromo, although these constructions have not been explicitly examined in pragmatic contexts. In Owens (1982:61) the postpositions of Boraana are categorized as variable, that is, found in genitival constructions as well a postnominally, and invariable, only occurring as postpositions. There are some differences in the inventories of these lexemes in Boraana and Guji. Owens identifies gubba "above?, dhaatu "next to?, dura "in front?, keesa "inside?, and duduuba "in back? as eligible as either nouns in the genitival construction and jiddu "between?, bira "with?, jal "under?, and irra "on top? as only postpositional. In Guji, gubbaa, dura, keessa, and duuba are always postpositions.

Owens (1982:61-63 and 1985:127) argues that these are best regarded as a subset of nouns. He identifies the markers found on postpositions as case markers. Since, however, when occurring as postpositions, they are eligible for morphological markers also found on postpositions and are restricted syntactically in the same ways that postpositions are, while no longer enjoying the privilege of co-occurrence of nouns, the recognition that they are both relational nouns and also members of the class of postpositions is more insightful. This reflects Hopper?s principle of layering, and also his principle of divergence (1991:24), which suggests that multiple forms with a common etymology function differently within the grammar of a language.

It seems that gara, discussed above in section 1.5, is transitional. It is fixed as a preposition in several dialects; it can only occur before the nominal in Guji. It does, however, occur in Waata, in a narrative from Stroomer (1987:241-242) after the noun, as in (63) line 2, where the site is already accessible to the audience.

(63) 1. yaanii d?eengedda gandiini k?uwatee, mala suu jaarti tokotti jira.

"Well, once upon a time a village settled (somewhere) and at that time there was an

old woman.?

2. d?iisa hink?abdu, ajoollee hink?abdi, hiyeeciaa; wara kaa gara teeci.

"She had no husband, no children, she was without relatives; she lived in this


However, it is found before the noun in (64) line 4, when it functions to identify a new site.

(64) 4. temteecee jaatiini suu taa kaatee guyya tokko, k?oraani d?ece, gara baddaa.

"She lived like this a long time, (until) one day she went to collect firewood in the bush.?

As is the case with relational nouns in Guji, in Waata, gara is found as a preposition where the site or path is novel, but postnominally where the site has already been established in the discourse.

4 Conclusion

This study has revealed the rich inventory of Oromo locational structures. Sites or paths may be signaled with postpositions, affixes, directionals, genitival constructions, the preposition gara, or some combination of these. Although there are a few unique locationals in some dialects, and some locationals are more grammaticalized in one variety or another, the core of locational elements and constructions is shared across Oromo speech communities. Morphological marking for path with a long final vowel on postpositions, affixes, directionals, and the prepositional phrase is found across dialects. Accessibility of referents or information that is signaled with the phrase-final n pragmeme across varieties is attested on locationals in Guji Oromo.

Grammaticalization has been observed within the synchronic grammar of Oromo. Hopper?s principle of layering is illustrated in various locationals as they range across the cline that has been proposed by Lehmann: the relational nouns, such as afaafa, that are more grammaticalized as postpositions, and the particles itti and irra that occur as the more grammaticalized affixes -tti and -rra. Lehmann?s notion of the relativity of autonomy and bondedness in grammar are also evidenced in the reduction of the affixed forms and their obligatory attachment to the preceding nouns. The principle of de-categorialization is realized in restrictions on co-occurrence, as when afaan functioning as a noun can be modified with a demonstrative, but not when functioning as a postposition. The principles of divergence and persistence are observed where verb particles itti and irra, and gadi and ol retain similar semantic values when they are grammaticalized as affixes or operate as directionals, and when they occur in frozen verb particle constructions and where gara is prepositional or reduplicated in idiomatic expressions.

The importance of the pragmatic status of information in motivating grammatical structure of locationals is demonstrated in the data by the choices of ordering that may be made by a speaker within a particular context depending on the status of the information available to the interlocutors at that point of the discourse. The relational nouns in genitive constructions are chosen when locations are not assumed to be accessible to all interlocutors, but the postpositions are selected when locations can already be identified by the discourse participants. The appearance of the n pragmeme on locationals in Guji is also determined by the assumed status of referents and information.

1 We thank Gerald Sanders, Tucker Childs, and three anonymous reviewers for their generous comments and suggestions that helped us in our work and Mekonnen Abakore for his native intuitions. All errors are ours alone.

2 This terminology varies among researchers. For example, Svorou (2002) uses trajectory to refer to the entity to be located, landmark to refer to the entity with respect to which it is being located, and relator to indicate the relation between them. Talmy himself states (1978: 628 fn 1) that he is not committed to the terms figure and ground, and he offers variable element and reference element as possible alternatives.

3 We use two separate terms for site and path because this difference is overtly marked in our Oromo data. Talmy uses path for both and (1978:641) an additional notion, State-of-Motion, to make this distinction.

4 The data examples drawn from the narrative in the Appendix in Yaachis and Clamons (2009: 182-192) are indexed in this paper parenthetically to the right of the examples. The data throughout are represented phonemically, following the conventions: Long vowels and consonants are represented with double letters or digraphs, the intermediate length vowel with an acute accent, the dental ejective with x, the retroflex implosive is represented with dh, the alveopalatal fricative with sh, alveopalatal affricate with c, the alveopalatal affricate ejective with ch, the velar ejective with q, and the palatal nasal with ny. Other abbreviations: PATH path, SUB subject, OBJ object, SEP separator, TOP topic, INDEF indefinite, GEN genitive, S singular, PL plural, 1-2-3 first-second-third person.

5 The final -f is glossed as "self? on the basis of contextual intuition. Gragg (1 976:183, 194) analyses -f as a dative case marker, and Stroomer (1995:112-113) as a benefactive on the verb. Bender, Eteffa, and Stinson (1976:140) identify this -f more generally as a particle with multiple functions that is found on nouns and verbs. Further study of -f is indicated.

6 When the -n , discussed in Yaachis and Clamons (2009) and in section 3.1 below, is attached to any stem, the final vowel is always lengthened.

7 Akka "like? also is prepositional in Guji, but since it is not locational we do not discuss it here. In other dialects, hamma "until? and others are also attested as prepositions (Gragg 1982: 226, Ali and Zaborski 1990:18) but are not used in Guji.

8 Baker and Kramer (2010:2) note that "OV + prepositions is the rarest kind of mix ed word order, found in only 10 of 1033 languages surveyed in WALS [the World Atlas of Language Structures]." The language samples considered by Heimstead (1992) and Svorou (1994, 2002) do not include any instances of a language utilizing both postpositions and prepositions in locational structures.

9 For this Harar Oromo story we are grateful to Roukiya Saad and Amal Osman.



Ali, Mohammed & Andrzej Zaborski. 1990. Handbook of the Oromo language. Wroclaw: Polska Akademia Nauk.

Baker, Mark & Ruth Kramer. 2010. "Prepositions" as Case Morphemes Inserted at PF in Amharic. Brussels Conference on Generative Linguistics 5: Case at the Interfaces.

Bender, Marvin Lionel, Mulugeta Eteffa and LLoyd D. Stinson. 1976. Two Cushitic languages. In Language in Ethiopia. M. L. Bender, J. D. Bowen, R. L. Cooper and C. A. Ferguson, eds. 130-154. London: Oxford University Press.

Clamons, Cynthia R., Ann E. Mulkern & Gerald Sanders. 1991. Case and topic marking in Oromo. Presented at the Annual Meeting of the Linguistic Society of America. 3-6 January. Chicago, Illinois.

Clamons, Cynthia R., Ann E. Mulkern & Gerald Sanders. 1993. Salience signaling in Oromo. Journal of Pragmatics 19. 519-536.

Gragg, Gene B. 1976. Oromo of Wellega. The non-Semitic languages of Ethiopia. M. L. Bender, ed. 166-195. East Lansing, Michigan: African Studies Center, Michigan State University.

Gragg, Gene B. 1982. Oromo dictionary. East Lansing, Michigan: African Studies Center, Michigan State University.

Griefenow-Mewis, Catherine. 2001. A Grammatical Sketch of Written Oromo. Cologne: R?diger K?ppe Verlag.

Griefenow-Mewis, Catherine & Tamene Bitima. 1994. Lehrbuch des Oromo. Cologne: R?diger K?ppe Verlag.

Heimstead, Lenief. 1992. Cross-language studies of adpositional marking form and function. PhD dissertation, University of Minnesota

Heine, Bernd & Mechthild Reh. 1984. Grammaticalization and Reanalysis in African Languages. Hamburg: Helmut Buske Verlag.

Hopper, Paul. 1991. On some principles of grammaticalization, in Elizabeth Traugott and Bernd Heine (eds.), Approaches to Grammaticalization. Amsterdam: John Benjamins. 17- 35.

Hopper, Paul J. & Elizabeth Closs Traugott. 1993. Grammaticalization. NY: Cambridge University Press.

Lehmann, Christian. 1985. Grammaticalization: Synchronic Variation and Diachronic Change. Lingua e stile, XX (3), 303-18.

Lehmann, Christian. 1995 (1982). Thoughts on Grammaticalization. LINCOM Studies in Theoretical Linguistics 01. M?nchen: LINCOM EUROPA.

Mey, Jacob L. 2001. Pragmatics: An Introduction. 2nd ed. Malden,MA: Blackwell Publishers.

Moreno, Martino Mario. 1939. Grammatica Teorico-Pratica della lingua Galla. Milan: Mondadori.

Owens, Jonathan. 1980. Observations on tone in the Booran dialect of Oromo (Galla). African Language Studies 17. 141-196.

Owens, Jonathan. 1982. Case in the Booran Dialect of Oromo. Afrika und ?bersee. LXV. 43-75.

Owens, Jonathan. 1985. A Grammar of Harar Oromo (Northeastern Ethiopia). Hamburg: Helmut Buske Verlag.

Stroomer, Harry. 1987. A comparative study of three Southern Oromo dialects of Kenya: phonology, morphology, and vocabulary. Hamburg: Helmut Buske Verlag.

Stroomer, Harry. 1995. A Grammar of Boraana Oromo (Kenya). Cologne: R?diger K?ppe Verlag.

Svorou, Soteria. 1994. The Grammar of Space. Amsterdam: John Benjamins.

Svorou, Soteria. 2002. Semantic constraints in the grammaticalization of locative constructions, in Ilse Wischer and Gabriele Diewald (eds.), New Reflections on Grammaticalization. Amsterdam/Philadelphia: John Benjamins. 121-42.

Talmy, Leonard. 1975. Semantics and syntax of motion, in Language Universals. 4. Syntax and Semantics. 181-238.

Talmy, Leonard. 1978. Figure and ground in complex sentences, in Joseph Greenberg (ed.), Universals of Human Language, Vol. 4, Syntax. Stanford: Stanford University Press. 625-649.

Traugott, Elizabeth Close. 2005. Constructions in Grammaticalization, in Janda, Richard D. and Brian D. Joseph (eds.). A Handbook of Historical Linguistics. Oxford: Blackwell. 624-47.

Yaachis, Mi?eessaa. nd. ms. Gospel of Mark.

Yaachis, Mi?eessaa & Robbin Clamons. 2009. Symbolic Indexing in Oromo. Studies in African Linguistics Volume 38. 2. 155-192.

[Author Affiliation]

Mi?eessaa Yaachis, Lenief Heimstead, & Robbin Clamons

Metropolitan State University

Mi?eessaa Yaachis, Lenief Heimstead & Robbin Clamons

Metropolitan State University

Dept of Literature and Language

700 East Seventh Street

Saint Paul, Minnesota 55106-5000

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.