Amharic is a Semitic, Afro-asiatic, language spoken mainly in Ethiopia, Northeast Africa by a significant number of the population of the country. It is the working language of the federal state. The language has also a fairly extended literary history and it functions as a medium of instruction, language of religions and language of the media (Aberra, 1997). This study discusses the polysemy of fit “face,” the Amharic body– part term (BPT) and its grammaticalized meanings. Grammaticalization is defined as “a process whereby expressions for concrete (= source) meanings are used in specific  contexts  for   encoding   grammatical   (= target) meanings” (Heine, 2003). Specific context may mean “highly specifiable morphosyntactic contexts, and specifiable pragmatic conditions or constructions” (Traugott, 2003: 625). In these specified constructions – collocated morphosyntactic strings fit “face” slots in and forms target meanings of grammatical nature.

The study discusses the grammaticalization of fit “face” from a BPT to spatial and temporal markers. The discussion focuses on seven spatial and temporal meaning extensions to front, side, parallel (opposite), future, past (before), temporal relations and negative past grammatical markers. The study  also  identifies  Amharicas having an “object–deictic orientation language” also known as “a single file model” (Heine, 1997). It means that Amharic uses a projective location marking “inherent to the reference object” (Frawley, 1992). In other words, Amharic spatial location using fit “face” and others are expressed based on the location of the object, rather than the speaker’s or the addressee’s position in the location.

The data has been collected from a wide-array of written sources and the internet due to absence of publicly available corpus data. The author of this study as a speaker of the language considered the data as everyday language and natural, not constructed for any specific purposes and not controversial. It is his belief that other speakers of the language will also consider them authentic.

Properties of grammaticalization

The study discusses the diverse extensions of meanings of fit “face” with the wide array of constructions following mainly functional grammaticalization models (Hopper, 1991; Heine, 1997, 2003; Croft, 2001, 2003; Traugott, 2003; Brinton and Traugott, 2005).

Grammaticalization has been conceived as a “historical change that results in the production of new functional forms” and the input to the process ranges from collocated strings, constructions, lexical items to grammatical items (Brinton and Traugott 2005). Moreover, mechanisms of grammaticalization are interrelated: these mechanisms in identifying the grammaticalization process are itemized as:

1. Desemanticization (loss in lexical meaning content)

2. Extension (use in new contexts)

3. Decategorialization (loss in morphosyntactic properties characteristic of the source forms) and

4. Erosion (or phonetic reduction) (Hopper, 1991; Heine, 2003).

Grammaticalization has stages (Heine, 2003): The initial stage where there is a linguistic expression “A” that is recruited for grammaticalization roughly corresponding to the use of lexical items and periphrastic constructions for functional purposes. The intermediate stage where the expression acquires a second use pattern, “B,” with the effect that there is ambiguity between “A” and “B” – roughly corresponding to periphrastic constructions and clitics, and the advanced–final stage where “A” is lost, that is, there is now only B (the source is difficult to be identified) which roughly corresponds to affixation and stem internal changes observed in human languages. At the advanced–final stage, grammaticalized items may show some or all of the following structural properties:

1. Paradigmatization, forming paradigms;

 2. Obligatorification, optional forms become obligatory;

3. Condensation, shortening of forms;

4. Coalescence, collapsing together of adjacent forms; and

5. Fixation, free linear orders become fixed (Hopper, 1991; Heine, 2003).

Based on the clustering of the meanings, the researcher argues here that BPTs are the sources/bases of spatial (section 3) and temporal (section 4) meaning extensions. Moreover, the meaning extensions from fit “face” to spatial and temporal meanings are not idiosyncratic. For instance, the set of Amharic effluvia terms, as adjectival constructions, encode personality traits (Aberra, 2014). There are well founded connections between the source and the target meanings. We have to be aware that, not every BPT will be the source for the location front, up, top, etc. The BPT has to be a front and upper part of the body. “There is a general process whereby certain body parts, on account of their relative location, are used as structural templates to express deictic location” of similar sort (Heine and Kuteva, 2002). On account of their position, shape and function, BPTs are also used in constructions, structural templates, “to anchor the spatial and temporal situation of the speech or the communication event” (Heine and Kuteva, 2002). BPTs in construction is also conveniently called “schema” (Croft, 2001). In their sample of over 400 world languages, Heine and Kuteva (2002) noted that “face” terms frequently grammaticalize to FACE (body part) > (1) FRONT, > (2) UP, TOP, ON, AT, and > (3) TEMPORAL. Amharic also uses fit “face” as spatial as well as temporal meanings.

Spatial meanings extensions of fit “face”

Grammaticalization of spatial grams is partially unique in Amharic from what can be observed in Heine and Kuteva (2002) or Hebrew (Petruck, 1986) studies, Amharic fit “face” does not pass through a stage in which it means TOP, UP, or ON instead, Amharic uses ras “head” and anat “temple of head” for these purposes: fit ‘face” is associated with the three interrelated spatial meanings; front, side (generic), contrasting side in Amharic. For instance, let us consider the following three Amharic constructions:

(1a)

fit–u                  ga

face–DEF         at

“at the front”

(Lit. “at his face”)

(1b)

fit–lÉ™–fit

face–PART–face

“facing each other opposite” or “at the forefront”

(Lit.‟face–to–face”)        (Leslau, 1995: 855; Gizaw, 2002: 708)

(1c)

wÉ™dÉ™–fitmÉ™t’t’–a

ALL–face come: PERF–3SG: M

“he came to the front”

(Lit. “he came to the face”)

In examples 1a and 1c, fit “face” collocates with adpositions ga “at” and wÉ™dÉ™ “to,” and forms adpostional phrases, whereas in example 1b fit “face” is totally reduplicated with the participle marker lÉ™ having an adverbial function. Two of the three examples (1a and 1b) have ambiguous senses between the intended meanings and the literal translations. As noted in the literal translation, 1a has also BPT readings. Hence the researcher listed a full context to see that they are associated with the spatial meanings as in 2 and as in 3 and 5.

“face” BODY PART > SPATIAL LOCATION > FRONT

In the following expressions – the fit “face” > has a meaning “front.”

(2)       

zaf–u    biro–u   fit–fit–ga tÉ™–tÉ™kkÉ™l–É™

tree–DEF          office–DEF       face–face–at     PASS–plant–3SG:M: SBJ

“the tree was planted in front of the office”

In this construction, fit “face” a BPT collocates with the inanimate object biro “office” and locates the position of the object, tree, by considering the office as a point of comparison. Moreover, as a locative postposition ga is optionally attached to the noun fit–fit as an alternative expression in example 2 but not in example 1a. The replication of fit in 2 as fit–fit encodes repetitive action of planting trees in front of each office. The expressions in 1b to c can also be disambiguated with similar constructions involving inanimate objects which do not have the body-part “face” meaning as such:

(3)       

biro–u   fit–lÉ™–fit lela biro all–É™

office–DEF       face–PART–face other   office   exist: PERF–3SG:M: SBJ

“there is another office opposite facing the office”

(Lit. “another office exists ‘face–to–face’ to the office”)

In this example, fit of the biro implies the front side of the office where there is an entrance and possibly windows similar to the actual  face  where  ears,  eyes,  nose,  and mouth where most entrance points to our body are located.

(4)

dan'el   fit–lÉ™–fit  hed–É™

Daniel face–PART–face go: PERF–3SG:M

“Daniel went to the fore front ”SPECIFIC SPATIAL MARKER

(Lit. “Daniel went face–to–face”)

fit–lÉ™–fit in the above example has a specific spatial meaning “front,” the construction has no sign of or need of entrance points.

(5)       

wÉ™dÉ™ hɨns’a–u    fit mÉ™t’t’–a

ALL building–DEF face  come:PERF–3SG:M

“he came to the front of the building.”

(Lit. “he came to the face of the building”)

In example 5 wÉ™dÉ™ “an allative marker” with fit “face” involves as a destination marker but the construction has entrance points.

(6)

dan'el   wÉ™dÉ™– fit hed–É™

Daniel ALL–face go: PERF–3SG: M

“Daniel went to the front” DIRECTION MARKER

(Lit. “Daniel went to face”)

This is the case in 5 and 6 where fit “face” and wÉ™dÉ™ “ the allative marker” co-occur with verbs of movement be it mÉ™t’t’a “he came” or hedÉ™ “he went.” Although example 5 has some sense of entrance due to the nature of the noun it modifies that is, hɨns’a “building,” in examples 6 and 7 there are no reference nouns such as biro “office” or hɨns’a “building” that signifies of any link between the body-part and extended meanings of fit “face.”

(7)

dan'el 'ɨ–fit hed–É™

Daniel LOC–face go: PERF–3SG:M

“Daniel went somewhere to the front”

(Lit. “Daniel went to face”) GENERIC SPATIAL MARKER

In this context, there is meaning change from something concrete to that of abstract – location marking. fit “face” occurs now in a wider context including inanimate objects which do not have any body part designations or senses.

“face” BODY PART > LOCATION > SIDE

fit “face” grammaticalizes to mean “side” as in the following context:

(8)       

šÉ™kɨm–u wÉ™dÉ™–and fit a–gadÉ™l–É™

load–DEF ALL–one face AGENT–weigh:PERF–3SG:M

“the load weighed to one side,” “it sided to one” (Gizaw, 2002: 708) SIDE MARKER

wÉ™dÉ™ and fit “to one side” (Lit. “to one face”) in example 8 shares the same schematic structure with equally possible constructions wÉ™dÉ™ k’əňň agadÉ™lÉ™ or wÉ™dÉ™ gɨra agadÉ™lÉ™ “it weighed to the right” or “it weighed to the left.” k’əňň and gɨra are “right” and “left” respectively. However, this expression of “face” as “side” is also used in contexts where one can compare the balance of other more abstract inanimate object or in the context of comparison where an issue from two sides weigh for pros and cons (9).

(9)       

kÉ™–hulÉ™t fit yÉ™–bÉ™ll–a daňňa

from– two face  RELZR–eat:PERF–3SG:M judge

“a judge who took bribe from the two (litigant) parties/sides”       BOTH SIDES MARKER

“face” BODY PART > LOCATION > PARALLEL

fit “face,” a BPT, grammaticalizes in its reduplicated form and collocates with the obligatory particle to mean “opposite side” as in 10.

(10)

fit–lÉ™–fit tÉ™–yayy–u

face–PART–face RECP–see: PERF–3PL

“they faced each other they confronted each other” (Lit. “the see each other face–to–face”)

The phrase fit–lÉ™–fit shared a syntactic schema with another locative expression but with contrasted meaning gon–lÉ™–gon “side by side” as in (11).

(11)

gon–lÉ™–gon tÉ™–k’É™mmÉ™t’–u

side–PART–side PASS–sit: PERF–3PL:SBJ

“they sat next to each other”

Instead of gon–lÉ™–gon “side by side” if we insert in the same construction fit–lÉ™–fit “face-to face/ opposite” the meaning of the whole construction changes from “next to each other” to “opposite to each other” or “facing each other.” In Amharic as in other languages, front and back side of the inanimate objects are identified, following various clues including presence and absence of entrance and visibility. For instance, buildings, offices and cars have identified anterior and posterior parts like the existence of doors, windows, gardens, etc. The use of fit “face” helps to identify anterior parts of these objects as exemplified in (12).

(12)

mÉ™kina–u  fit lɨj–u k’omm–É™

car–DEF face boy–DEF            stand: PERF–3SG:M:SBJ

“the boy stood in front of the car”

It is safe to conclude that Amharic does not involve the speaker in locating the boy, the object or the car in 12. The location is from the view point of the object, not even the possible !lɨj–u huala mÉ™kina–u k’omm–É™ “the car stood behind the boy” is used as a prototypical locative marking. In the words of Frawley, this projective location marker made Amharic as having an “inherent to the reference object” parameter as against to “an inherent to the viewer” parameter (Frawley, 1992). Similarly, if we consider Heine’s (1997) basic systems of spatial orientation or reference, Amharic has an “object deictic orientation” system.  Let us consider, inanimate objects which do not have designated front and back, for instance, tree and ball using Frawley’s (1992) schematic situation and of course the viewer in order of – the face of the viewer-tree-ball to answer the question: what spatial expressions of tree and ball does Amharic provide?

(13)      

kuas–u  kÉ™–zaf–u fit nÉ™–w

ball–DEF from–tree–DEF face be: PERF–3SG: M

“the ball is in front of the tree”

Amharic uses the side of the tree facing the ball as front, the language projects the viewer’s direction of the tree. Objects inherit the vantage point of the speaker or the viewer. The same vantage point works to the ball. It is possible to say in Amharic as in example 14a and b.

(14a)    

zaf–u kÉ™–kuas–u jÉ™rba nÉ™–w

tree–DEF from–ball–DEF back  be:PERF–3SG:M

“the tree is at the back of the ball”

(14b)    

zaf–u kÉ™–dan’el  fit nÉ™–w

tree–DEF from–daniel–DEF face be:PERF–3SG:M

“the tree is in front of the Daniel”

In this regard, Hausa, a Chadic Afroasiatic language, has a different view of the ball which is not between the speaker and the tree (as in 13) from Amharic.

(15)

go kwallo can baya da itace

look ball there back with tree

“there is the ball behind the tree” (Frawley, 1992, p. 269).

In the Hausa sentence, the face of the tree is facing the speaker whereas in Amharic the tree’s back faces the speaker. The Hausa case is known as “a face–to–face model” of a relativistic perspective (Heine, 1997). To conclude, using  fit  “face”  a  BPT  as  a  spatial  marking

Amharic identifies both projective location marker – as INHERENT TO THE REFERENCE OBJECT (OBJECT–DEICTIC ORIENTATION) and following a SINGLE–FILE model. In a typical grammaticalized meaning fit “face” occurs in its bare form without noun inflections but with adpositions. In all instances of the spatial marking fit “face” shows the obligatorification tendency co–occurring with adpositions, the usual allative preposition wÉ™dÉ™ “to” and the locative postposition ga “at.” In the following section, temporal meaning associated with fit “face” shares the same schema a la  Croft (2001) [prep– face][postp] with spatial meaning and the lexically specified allative preposition wÉ™dÉ™ “to.”

The temporal marking of fit “face” in addition to sharing the same schema and the same lexically specified allative preposition, it uses a plethora of other prepositions such as kÉ™ “from,” bÉ™ and yÉ™. Hence, their presence in the schema made possible for fit “face” to encode several time contrasts future–past, contrasting temporal relations, ordering in time (chronology), and negative past in a variety of constructions.

 

“face” BODY PART > TEMPORAL > FUTURE

 

(16)      

tÉ™mari–u wÉ™dÉ™–fit yɨ–mÉ™t’t’–al

student–DEF     ALL–face IMPF–come –AUX: 3SG: M

“the student will come”

(Lit. “the student is coming to face”)

 

“face” BODY PART > TEMPORAL > PAST/ BEFORE

 

(17)

dan'el   bÉ™–fit     hed–É™

Daniel at–face go: PERF–3SG: M

“Daniel went before,” TEMPORAL MARKER

(Lit. “Daniel went before”)

 

(18)      

bÉ™–fit yÉ™–mÉ™t’t’–a–u sÉ™w yÉ™t–u nÉ™–w

ABL–face RELZ–come: PERF–3SG: M   man which–def be: PERF–3SG: M

“who is the person who came early”

(Lit. “which is the person who came by face”)

 

Example 18 has the meaning sense of “earlier than all” or “who came before all the others came.” Example 18 is also a good instance of the meaning “before” contrasting with example (19), an example of “after”.

 

(19)      

bÉ™–huala yÉ™–mÉ™t’t’–a–u  sÉ™w      yÉ™t–u nÉ™–w

ABL–back RELZ–come:PERF–3SG:M man which–DEF

be: PERF–3SG:M

 

“who is the person who came late at the end”

(Lit. “which is the person who came at back”)

 

bÉ™–fit as a full-fledged construction slots in where opposing body-parts can fill in to render contrasting meanings. In this sense, 20a and 20b contrasts each other. In other words, fit and huala form paradigms that is, they have similarity by being body-part terms, filling in the same construction slot adposition+N and collocates with either allatives or ablatives. These examples of the meaning extensions of the fit “face” as temporal markers together with other examples in spatial marking indicate that fit “face” is between intermediate and advanced stages of the grammaticalization processes depending on the semantic and the syntactic criteria discussed in section 2.

 

(20a)    

kÉ™–fit yÉ™–mÉ™t’t’–a guday nÉ™–w

ABL–face RELZ–come: PERF–3SG: M   issue be: PERF–3SG: M

“the issue came from the past.”

(Lit. “that was the issue that came from the face”)

 

(20b)    

kÉ™–huala yÉ™–mÉ™t’t’–a guday        nÉ™–w

ABL–back RELZ–come: PERF–3SG: M  issue be: PERF–3SG: M

“the issue that surfaced later in the process.”

(Lit. “that was the issue that came from back”)

 

In the above four examples 18 and 19, 20a and 20b contrast with each other and gave different time senses. The query, how come that fit “face” a single body-part term has signalled opposite temporal meaning, it is neither fit “face,” nor the ablative – the allative particles that mark the temporal time relationship, but it is the whole constructions that expresses the opposite intended meanings as noted in the above four examples.

 

“Face” body part > temporal relations

 

Fit “face” with the associated meaning “before” collocates with huala “back”> “later/after” to encode contrasting temporal-relation as observed in some historical texts of Amharic of the 1800 century.

(21)      

kosso–na kɨrÉ™mt  fit yɨ–mÉ™rral huala gin dÉ™s yasÉ™ññal

kosso–and winter face   bitter:IMPF:AUX later but happy make:IMPF:AUX

“kosso and winter are bitter before, but later they will make someone happy.”

(22)

ɨndÉ™–zih yallÉ™ moññ yÉ™–fit–u–n ɨnji  yÉ™–huala–u–n

like–this exist  fool PART–face–DEF–ACC but PART–back–DEF–ACC

yəmayay

see:IMPF:NEG:3SG:M

“A fool like this who sees what is now but does no see/think about the future.”

Like temporal contrast relations the same paired BPTs fit–huala expressions make parallel spatial front– back contrast as in (23).

(23)      

nɨgus    tewodros –m kÉ™–fit–É™ñña–u sÉ™lf  k’omm–É™u

king      theodore–REFL ABL–face–ADJ–DEF     parade  stand: PERF: 3SG: POL

kÉ™–hual– É™ñña–u sÉ™lf bɨzu fÉ™rÉ™s–É™ñña a–zur–É™u

ABL–back–ADJ–DEF  parade  many cavalry AGENT–return: PERF–3SG: POL

sÉ™dÉ™d–u–bb–É™t

send: PERF–3SG: POL –MALF–3SG: OBJ

“King Theodore, standing at the front parade, sent a cavalry to the back parade.”

In both, the temporal contrast relations and the spatial front–back contrast, we observe the collocation of fit–huala “face–back” with the same particles, definiteness and in contrasting clausal schema.

“Face” body part > negative past

In the following schema, bÉ™–noun–TEMP–DEF–ACC–u with the suffix –u has several different meanings from negation to that of the focus marker: fit “face” occurs associated with the negative past meaning as in (24).

(24)

bÉ™–fit–u–n–u bɨ–tɨ–hed    t’ɨru nÉ™bbÉ™r

PART–face–DEF–ACC–NEG       if–2SG: M–go  good be: PERF

“it was good, if you had gone before”

Compare example (24) with the contrasting sense, that is, present time ahun “now”, but sharing the same construction schema with as in 25.

(25)

ahun–u–n–u bɨ–tɨ–hed     t’ɨru nÉ™w

now–DEF–ACC–NEG    if–2SG: M–go   good    be: IMPF

“it is good, if you go right now”

Lastly, fit “face” is also associated with the temporal ordering as in the following expression.

(26)

yÉ™–fit–É™ñña–u hulÉ™tÉ™ñña–u–n k’É™dɨm–o    yɨgÉ™bal

PART–face–ADJ–DEF two–ADJ–DEF–ACC precede–GERUND enter: IMPF: AUX

“the first one entered preceding the second one”

Moreover, as Heine and Kuteva (2002) identified that, face, a BPT grammaticalizes into TEMPORAL gram meaning “before,” Amharic grammaticalizes fit “face” to fit “before”. Moreover, when collocating with various adpositions such as the ablative kÉ™ “from” bÉ™ “by/from” or an allative wÉ™dÉ™ “to” markers, fit “face” grammaticalizes into “past” and “future” markers.

This study focuses on meaning extensions associated with the Amharic body-part term, fit “face.” The study identifies interesting extensions of the grammatical meanings of fit “face”; three spatial (front, side, parallel / opposite), and four temporal (future, past, temporal relations and negative past) with co–occurring adpositions. The adpositions except the well identified goal marker, the allative wÉ™dÉ™, and the source marker ablative kÉ™, all the others like yÉ™, bÉ™, and lÉ™ are not identified for their possible meanings and distributions. The temporal suffix –u is also another case that awaits further comprehensive study.

Based on the ambiguous meaning extensions in the spatial domain between its literal and its grammaticalized meanings fit “face” can be designated as in the intermediate stage of the grammaticalization process and based on fulfilling three of the five syntactic criteria (forming paradigms, become obligatory in the constructions and fixed in its linear order), fit “face” is in the advanced-final stages of the grammaticalization processes.

Moreover, the study of fit “face" as a BPT and its associated extended meanings help to identify two issues about the spatial basic system of Amharic – object–deictic oriented, inherent reference object of Amharic which may not be identified using other possible lexical items: Amharic in its projective spatial expression, the viewer projects and imposes its vantage point to the inanimate objects. Heine (1997) calls this projection a single –file model. Comparing expressions in the two Afroasiatic languages of Africa (Amharic and Hausa (14a) and (15)), the two languages, Amharic and Hausa seem to follow different models– single–face–model and face–to–face model respectively. However, Heine (1997) recalls Hausa too having a single–face model like Amharic. This issue raises the question either the models do not seem to work for Amharic or the Amharic analysis may be wrong. Still by having the difference of the model and the projection identified for Amharic tree or using Heine’s model the box, both Frawley’s and Heine’s schemas are about the ball and the mountain which is at the far end in the order. All these require further  research on the issues.

Lastly, a quick comparison of the grammaticalization paths of fit “face” in Amharic and panim “face” in Hebrew (Petruck, 1986), they are both genetically-related to Semitic languages of the Afroasiatic phylum but spoken at different locations, gave partially different results. Hebrew panim “face” grammaticalizes in to “before,”  “against,” and “because” (Petruck, 1986), whereas Amharic fit “face” does not have the extended meaning sense of “against” or “because” at all. It would be interesting to perform a grammaticalization study of “face” in other Cushitic, Omotic, and Nilo–Saharan languages of Ethiopia to see if they share the same grammaticalization paths with Amharic as they share the same geographical area. Then it is possible to hypothesize that due to language contact situation, languages in the same geographical regions may share the same grammaticalization paths, regardless of their genetic relationship.

The author has not declared any conflict of interests.