Volume 16 Issue 1 (2018) DOI:10.1349/PS1.1537-0852.A.486 Note: Linguistic Discovery uses Unicode characters to represent phonetic symbols. Please see Optimizing Display for requirements to accurately reproduce this page. Comparison in Kambaata: Superiority, Equality and
Similarity
Yvonne Treis CNRS-LLACAN This paper is an in-depth study of the expression of
comparison in Kambaata, a Highland East Cushitic language of Ethiopia. It
discusses not only quantitative comparison, i.e. comparison of relative and
absolute inequality and comparison of equality, but also analyses the morphology
and syntax of expressions of qualitative comparison, i.e. comparison of
similarity. Apart from predicative constructions, the analysis takes into
account attributive comparative, superlative, equative and similative
constructions. In the comparative construction (lit. ‘X is tall from Y’), the
standard of comparison is marked by the ablative case, as in most languages
spoken in the Horn of Africa. Kambaata distinguishes between two superlative
constructions, one of which is based on the comparative construction (‘X is
tall from all’), while the other is characterised by a locative standard of
comparison (‘X is tall among Y’). Furthermore, Kambaata has two equative
constructions. The first is based on the similative construction (‘X is tall
like Y’); the second is a periphrastic construction (‘X is tall to the extent
Y’). The paper argues that the enclitic morpheme which marks the standard of
comparison in the similative construction originates from a noun meaning
‘manner’. 1. Introduction
This
paper is a study of the expression of comparison in Kambaata, a Highland East
Cushitic language of Ethiopia. It encompasses analyses of quantitative and qualitative
comparison.[1]
It is concerned with comparison of inequality (§3), similarity (§4) and equality
(§5). While works on comparison in little known languages often focus on predicative
comparison constructions with simple nominal standards of comparison, I will go
further, examining constructions with complex clausal standards and attributive
comparison constructions. The first sections of this article provide information
on the classification and location of the language (§2) and give an overview of
important grammatical aspects that are required as background for the discussion
of comparison (§3-5). To the best of my knowledge, this paper is so far the
most detailed study of the expression of comparison in a Cushitic language. My
analyses are based on data collected during fieldtrips to the Kambaata-speaking
area from 2002 to 2007 and in 2016. Apart from narratives, conversational data and
mock dialogues,[2] my corpus also
includes texts from local publications.[3] These sources are
supplemented by elicited data. I have generally attempted to avoid translation
elicitation but have resorted to target language manipulation elicitation or
text-based (ancilliary) elicitation (Chelliah & de Reuse 2011: 370-73, 379),
whenever elicitation was necessary at all. All data is presented in the
official Kambaata orthography (see Treis 2008: 73-80), supplemented by accents
to indicate phonemically distinctive stress. 2. Typological overview of Kambaata
Within
the Cushitic branch of the Afro-Asiatic language phylum, Kambaata belongs to
the Highland East Cushitic (HEC) language group. It is spoken in the South of
Ethiopia in an area approximately 300 km south-west of the Ethiopian capital
Addis Ababa in the Kambaata-Xambaaro Zone (Figure 1). The immediate neighbours
of the Kambaata are speakers of other Highland East Cushitic languages (Alaaba
and Hadiyya) and Ometo languages of the Omotic family (Wolaitta and Dawro). According
to the 2007 Ethiopian census, there are 615,000 Kambaata speakers.
Figure 1. Kambaata-Xambaaro Zone and its woredas (Map designed by Jérôme Picard, CNRS-LLACAN, 2016; Sources: ESRI, USGS
data.humdata.org; all boundaries are unofficial) Kambaata
is exclusively suffixing and, regarding its morphological type, agglutinating-inflectional
with many portmanteau morphemes. It is both head- and dependent-marking with a
fairly elaborate case system and subject agreement on verbs. It is consistently
head-final; hence all modifiers, including relative clauses, precede the noun
in the noun phrase, and all dependent clauses precede independent main clauses.
The main verb or a copula is usually the last constituent in the sentence. Clefting
is a very common focussing device. Kambaata has four major open word classes: nouns,
adjectives, verbs and ideophones, all of which can be defined on the basis of
morphological and morphosyntactic criteria (Treis 2008: 81-97). Nouns are
obligatorily specified for one of nine cases, and for either masculine or
feminine gender. Table 1 exemplifies the case paradigms of the masculine noun dum-á ‘back room (of a house)’ and the
feminine noun gat-í-ta ‘backyard’. The
accusative is the functionally unmarked case. It marks not only direct objects
but also certain temporal and manner adverbial phrases, and it serves as the
citation form of nouns.
Table 1. Case paradigms of a masculine and a feminine Kambaata noun In
attributive function, adjectives agree in case and gender with their head noun.
They are marked for three cases (nominative, accusative, oblique) and two
genders (masculine, feminine). Apart from being used as modifiers, adjectives
can be used as the head of a noun phrase without further measures (e.g. nominalisation,
dummy head) being taken. In this syntactic function, they have the full
case-marking potential, i.e. they inflect for nine cases. Kambaata makes a morphological distinction between fully
finite main clause verbs and various types of semi-finite and non-finite dependent
clause verbs: relative verbs, converbs, purposive verbs and verbal nouns. In Table
2, the verbs are arranged from left to right on a scale of decreasing
finiteness. While Kambaata has nine independent, case-inflecting personal pronouns
(1s, 2s, 2hon, 3m, 3f, 3hon, 1p, 2p, 3p), subject agreement morphemes on verbs
distinguish maximally between seven (1s, 2s, 2hon = 2p, 3m, 3f = 3p, 3hon, 1p),
and minimally between five forms (1s = 3m, 2s = 3f = 3p, 2hon = 2p, 3hon, 1p).
Table 2. Inflectional categories
on main and dependent verbs in Kambaata (+ / - = category can or cannot be morphologically marked) Ideophones
are invariable and are always accompanied by a support verb which carries the
inflection. In intransitive clauses ideophones combine with y- ‘say’ or ih- ‘become’, and in transitive clauses with a’- ~ ass- ‘do’. 3. Comparison of inequality
Comparison
of inequality subsumes comparison of superiority and inferiority. For both
types of comparison, a distinction can be made between relative inequality (e.g.
Susan is taller than Peter; Peter is less
tall than Susan) and absolute inequality
(superlativity) (e.g. Susan is the
tallest of her family; Peter is the least tall of his family). Kambaata does
not have a grammaticalised construction for comparison of inferiority,
but expresses inferiority periphrastically (§3.3). Hence the focus of
this section will be on comparison of superiority (§§3.1-3.2). 3.1.
Comparison of superiority
In
the Kambaata predicative comparative construction, the comparee functions as
the subject. As such, it is nominative-marked, e.g. Bóq, the nominative form of the personal name Boqé in (1). The parameter of comparison is expressed by the
predicate, which is a member either of the word class of adjectives (1), or of
the sub-word classes of property ideophones (2) or property verbs (4).[4] The comparee in
subject function triggers gender agreement on adjectival parameters and
person/gender agreement on verbal parameters and on support verbs of ideophonic
parameters (agreement is marked by underlining in (1)-(2)).
There
is no grammaticalised parameter marker; the form of the parameter itself is no
different from that of the positive construction (see the literal translation).
The standard of comparison, i.e. the entity against which the comparee is
measured and found to be unequal, is expressed by an ablative-marked adjunct.
In (1) and all following examples, the standard phrases occur in curly brackets
in the Kambaata and the translation line. The
primary elements of the predicative comparative construction and their
grammatical functions are summarised in (3). (3) Predicative
comparative construction
Whereas
two entities are compared in (1)-(2), one can also compare two conditions of
one and the same entity at different times. In (4), a current medical condition
is compared to the previous condition, and a temporal noun serves as the
standard of comparison.
The
occurrence of the ablative case is not restricted to the comparative construction.
Elsewhere in the language, the ablative case marks oblique objects and adjuncts
which may express the source (5), origin or starting point of an actual
(literal) or metaphorical (figurative) movement, the starting point in time,
the source material of a production process, what is avoided, what one is
protected or saved from (6), or the maleficiary of an event.
Furthermore,
some relational nouns, such as zakk-ú ‘after’, etar-ú ‘beyond, exterior;
apart (from)’ and bir-íta ‘before,
front’, and some adjectives, such as annann-á(-ta) ‘different (from)’ and wol-ú/-íta ‘other
(than)’ (7), govern ablative complements.
The
Kambaata canonical comparative construction would be categorised into the
common comparative typologies as follows: In Stassen’s typology, it would be
labelled a Separative Comparative (1985: 39f), because the standard of
comparison is encoded as an adverbial phrase with a separative (‘from’)
interpretation. Like most languages employing this comparative construction
type (Stassen 1985: 40), Kambaata has SOV constituent order. In Stassen’s
less-fine grained WALS typology (2013), the Kambaata comparative construction
would be labelled a Locational Comparative, the most common worldwide type. In
Dixon’s (2008) typology, the Kambaata comparative construction would be
labelled type A1, which subsumes constructions in which the parameter is the head
of a copula complement or a verbless clause complement and in which the
comparee is encoded as the subject and the standard of comparison as an oblique
NP (2008: 789f). In Heine’s (1997: 112) typology, the Kambaata comparative
construction follows the Source schema (‘X is Y from Z’). Zelealem & Heine
(2003: 56f) claim that the Source Schema is the primary schema of the Ethiopian
Linguistic Area but uncommon elsewhere in sub-Saharan Africa. Thus Kambaata is a typical Ethiopian language with respect
to the encoding of comparison of superiority. 3.1.1.
Constituent order
As
seen in (1)-(2) above, the basic constituent order in the comparative
construction is Comparee – Standard –
Parameter As
is expected of a rigidly head-final language, the parameter is always the final
constituent of the comparative construction. The order of comparee and standard
is, however, pragmatically determined to a certain extent, and examples in
which the comparee follows the standard are also attested in my corpus.
Interrogative comparative examples are a case in point: in (8), a speech act
participant is asked to which entity the parameter of comparison (richness) is
applied to a higher degree. By the very nature of being questioned, it is undetermined
which one of the two compared entities serves as the comparee
or standard. Hence both entities are encoded in coordinated, additive-marked
ablative NPs. The subject function is occupied by a question pronoun enquiring about the comparee.
The question pronoun is placed in the pre-predicate slot, as is typical of
non-clefted interrogative sentences in Kambaata. (Standard + Standard) – Wh-Comparee
– Parameter
If
the sentence focus is on the comparee it occurs in the pre-predicate slot.[6] Standard – Comparee[Focussed]
– Parameter
The
canonical constituent order Comparee – Standard – Parameter
is not retained in the attributive comparative construction (see §3.1.4). 3.1.2. Higher
and lower degree of superiority
The
comparative construction can be expanded by adverbial modifiers specifying the
degree of superiority. The converb form of abbis-
‘exceed’ (11),[7] multiplicative
numerals (12) or ACC-marked adjectives in adverbial function, e.g. qah-ú ‘small, a bit’ (13) and baas-á ‘much more’ (9), occur before the
parameter.
The
adjective abb-á(-ta) ‘big; much’ fulfils
various functions in the comparative construction. Firstly, it can express the
parameter of comparison as in (14).
Secondly,
if modifying the parameter, it marks a higher degree of superiority of the
comparee with respect to the standard (15), and thus has the same function as
the degree adverbials in (11)-(13).
If
fully reduplicated, abb-á ‘big; much’
can express an extra-high degree of superiority (16).
Thirdly,
abb-á ‘big; much’ makes comparison of
superiority possible, if the predicate of the clause is not an adjective, property
verb or property ideophone; see the non-property verb qorab- ‘take care’ in (17).
For
the use of abb-á ‘big; much’ with
clausal standards see §3.1.3. Kambaata has two inherently comparative adjectives, woyy-á(-ta) ‘better’ (24) and baas-á(-ta) ‘worse, (negative) more’ (19),[8] and two corresponding
inchoative-stative verbs woyy- ‘be(come)
better’ (18) and baas- ‘be(come)
worse, (negative) more’.[9] Even if no standard
of comparison is mentioned in the immediate context, these property lexemes imply
that two (or more) entities are compared to each other and that the comparee
exhibits the parameter encoded by the adjective to a higher (or, depending on
the context, to the highest) degree.
3.1.3.
Clausal standards
If
a whole clause headed by a finite verb expresses the standard, it is nominalised
(NMZ2) so that it can carry a case morpheme (20)-(21). Clausal standards are commonly
followed by the degree marker abb-á
‘big, much’ (20) or aluud-ú (ACC) (20)
/ alúud-iin (ICP) ‘above’ (21).[10]
If
the clause is headed by a verbal noun, which retains the argument structure of
a verb but inflects like a noun (Treis 2012), no additional nominalisation is
necessary. In (22), verbal nouns are used as heads of the standard and the
comparee NP.
3.1.4.
Attributive comparative construction
Apart
from the predicative comparative construction, Kambaata also has an attributive
comparative construction (23). Here all components of the comparative
construction are found inside one NP. The comparee is the head of the NP, which
can occupy any syntactic function in the clause. It is modified by a phrase
containing the parameter preceded by the ablative-marked standard. Adjectival
parameters agree with the comparee in case and gender (23-i); verbal parameters[12] show gender and person
agreement and need to be relativised (23-ii). If the relative verb expressing
the parameter is negative, it also agrees in case (and, again, in gender) with
the head;[13] this additional
agreement morpheme occurs in round brackets in (23-ii). (23)
Attributive comparative construction
The
attributive comparative construction is exemplified in (24)-(26). In (24), two
coordinated comparee nouns (‘metal and stone’) are modified by an adjective, expressing
the parameter (‘better’), which itself governs the adjunct expressing the standard
of comparison (‘from/than wood’) (cf. (23-i)). The comparee nouns in (24) are
marked for the nominative case because they function as the subject of dagámmee ‘(3m) is found’.
In
(25), the noun expressing the comparee (‘poem’) is the object of xáaf ‘Write!’ and requires the
accusative case in this function. The preceding modifier phrase contains a
negative relative verb expressing the parameter of comparison, which itself
governs the standard noun phrase (‘from/than ten lines’) as an adjunct.
The
attributive comparative construction becomes fairly complex when the standard
of comparison in the modifier phrase is again modified. In (26), the standard taméech ‘from/than the use’ is preceded by
the genitive noun daabb-ó ‘of bread’,
which is itself modified by two coordinated genitive nouns (‘of wheat and oats’)
and a relative clause (‘which is very common in Ethiopia’).
3.2.
Comparison of absolute superiority
For the expression of absolute
superiority (superlativity), we find two different superlative constructions in
Kambaata. One has an ablative-marked universal quantifier as the standard of
comparison; the other has a locative standard of comparison. Most commonly, the
standard phrase precedes the comparee phrase. The English paraphrases are as
follows: (27) (i)
Ablative superlative construction: ‘From all (X), A is good.’ These constructions are considered semantically
equivalent by native speakers; they may, however, have pragmatic differences that
remain to be explored. 3.2.1. Ablative
superlative construction
The ablative superlative construction is built on the model of that expressing
relative superiority (see §3.1); the only difference concerns the choice of the
standard of comparison, which is either the universal quantifier horá ‘all’ (28a) or a numeral quantifier
(28b) that specifies the number of members in the set with which the comparee
is compared. The individual standard of comparison of the comparative
construction is replaced by a standard denoting the entire set of possible
individuals in the superlative construction.
The quantifier can also be
used as a modifier in the standard phrase, as in (29)
where the head of the standard phrase is the nominalising
enclitic =r ‘ones, thing(s)’.
A morpheme -n of a yet undetermined pragmatic function
– and hence just glossed “N” – is obligatorily added phrase-finally after the
ablative morpheme of the standard NP in the superlative construction; cf.
(28)-(29). While this -n-morpheme is
also found in other contexts,[15] it is to be noted that the standard in the comparative
construction of relative superiority (§3.1) is not combined with this morpheme
except in the following grammatically determined context: If the standard of
comparison contains a disjunction, the conjunct preceding íkko obligatorily co-occurs with the -n-morpheme (30). Note, however, that the connector íkko triggers the occurrence of -n in all contexts, i.e. even outside of
the comparative construction. More work is still required on the functional
range of ‑n, which is simply labelled
an “emphasis marker” by Schneider-Blum (2007) in Alaaba, Kambaata’s closest
relative.
Kambaata’s primary
superlative construction, in which the standard is expressed by a universal
quantifier, belongs to the cross-linguistically most widespread Type A (Absolute
Comparison Superlatives) in Gorshenin’s typology (2012: 83). 3.2.2. Locative
superlative construction
Whereas the ablative superlative
construction (§3.2.1) is widely attested in my corpus, the locative type is
less common. Two instances are attested in a mock dialogue, on the basis of
which I was able to elicit several additional examples; two other examples were
overheard, and one example was found in a local publication. The ablative
superlative construction (§3.2.1) can be considered a sub-type of the comparative
construction (§3.1). In contrast, the locative superlative construction is not directly
formally related to the comparative. Instead, it is a simple positive
construction, i.e. non-comparative qualifying construction, extended by a
locative adjunct. The locative standard qophphan-áan
‘among lies’ in (31) refers to the totality of lies from which the comparee,
isé=bíi ‘hers (i.e. her lie)’, is
singled out and to which it is compared. Another superlative construction with
a locative standard is given in (32). The ablative and locative superlative types
do not only differ with respect to the case marking on the standard NP but the locative type also lacks a universal or numeral
quantifier.
A native speaker confirmed
that qophphanáan ‘among lies’ (31) and
minéen ‘among houses’ (32) could be
replaced by ablative standards plus a universal quantifier (§3.2.1), i.e. qophphaníichch
horíichchin ‘from/than lies’ and miníichch
horíichchin ‘from/than all houses’, respectively. Ex. (33)
is taken from a local Kambaata publication on oral literature; ex. (34) was
overheard.
Unlike the comparative construction (§3.1) and the
ablative superlative construction (§3.2.1), the locative superlative
construction has a rigid word order. As seen in all attested examples, the
standard of comparison always precedes the comparee. Permutations tests have
shown that the superlative interpretation is lost if the order of standard and
comparee is reversed (35)-(36).
Kambaata’s secondary
superlative construction, in which the scope is expressed in a locative
adverbial phrase and marked by a morpheme not used as a standard marker in the comparative
construction, belongs to Type S[cope] Superlative of Gorshenin’s (2012: 111f)
typology. 3.2.3.
Attributive superlative construction
Kambaata
has an attributive superlative construction in which all components of the
comparison (comparee, standard and parameter) are found inside one NP. In (37), the comparee kasalá ‘charcoal’ is the direct object of áffeemm ‘I have grabbed (i.e. I have)’. In the comparee NP, of
which kasalá is the head, we find a
modifier consisting of the adjective danaam-ú
‘good’, expressing the parameter of comparison, and the universal standard hor-íichch-in ‘from all’.
Adjectives can be used as heads of NPs without
being nominalised, as the use of fayy-á ‘healthy’
in (38) illustrates. The referent with the quality expressed by the adjective is
determined by the (extra-)linguistic context.
Hence, we also find examples in which the comparee
of an attributive superlative construction is retrievable from the context and thus
omitted. In (39), the adjective kee’mmáashsh-u
‘(the) heavy (ones)’ has become the head of the NP from which the comparee was
dropped.
Attributive superlative examples
with a locative standard NP (§3.2.2) are not (yet) attested in recorded or written data.
Elicitation confirmed, however, that they are possible (40).
3.3.
Comparison of inferiority
Kambaata
does not have a grammaticalised or conventionalised comparative construction of
inferiority. In order to express that someone has a quality to a lesser degree
than somebody else, periphrases with the inchoative-state verb kot- ‘be(come) not enough, less, insufficient,
small, decrease’ are possible but rarely attested in my database (41).
Instead, inferiority is expressed by negative
equative constructions (§5) (42)-(43).
4. Comparison of similarity
A similative construction is a type of comparison construction expressing equal manner or quality between a comparee and a standard of comparison (Haspelmath & Buchholz 1998, Fuchs 2014). Kambaata marks the standard of similative comparison by an enclitic multifunctional morpheme =g (gloss: G); the standard is a genitive modifier to =g (44).
The
standard phrase is an adverbial modifier to the predicate. It consists
minimally of a genitive noun plus =g;
the enclitic is itself case-marked; see -a
ACC/OBL in (44). The double case-marking in the standard phrase – once on the
semantic head, once on the standard marker =g
– points to a nominal origin of the enclitic. The standard marker =g is in fact a manner nominaliser.
Before proceeding with the analysis of the similative construction (§4.4),
information on the historically primary function of =g as a manner nominaliser is provided in §§4.1-4.3. Treis (2017b) gives a
more detailed account of the multifunctionality of =g, which is used, among others, as a marker of complement clauses (85),
purpose clauses and temporal clauses of immediate anteriority (‘as soon as’). 4.1.
The morphology of the manner nominaliser =g
The
morpheme =g belongs to a group of
enclitic nominalisers including =b ‘place’
(45), =bii(‑ta) NMZ1b ‘one (m/f)’
(31)-(32), =hann NMZ2 ‘one (m)’ (2), =tann NMZ2 ‘one (f)’ and =r NMZp ‘thing(s), ones’ (7), (29), all
of which are of (pro)nominal origin and take modifier phrases, i.e. genitive (pro)nouns,
inflected adjectives and relative clauses rather than roots or stems as their
input.[16]
The
case marker following the nominalisers is not fixed but dependent on the
syntactic function and the semantic role of the phrase in the clause, e.g. the
ACC/OBL-marking -a on the standard
phrase in the similative construction in (44) is due to its adverbial function.
The morpheme =g is inherently
masculine (see the gloss of the case/gender portmanteau suffix with which it
combines) and inflects almost like a full noun (Table 3).
Table
3. Case paradigm of =g compared to
that of masculine full nouns The
case paradigms of =g and the noun dum-á ‘back room’, a masculine noun of
the ‑á-declension, are almost
identical.[17]
The =g-morpheme only lacks a
difference between the accusative and oblique case (the syncretic form is glossed
ACC/OBL), which is a type of syncretism not attested for any other nominal
declension.[18] While the vast
majority of masculine nouns distinguish between eight cases, the =g-morpheme distinguishes only between
seven. Another small difference concerns the genitive form, which is a long ‑íi for the manner morpheme but a short ‑í for nouns of the ‑á-declension. The equal sign indicates that the enclitic =g is phonologically and syntactically dependent on a host. It is
stressless in certain cases and can never be used in isolation. 4.2.
The manner nominalising function of =g
The
=g-morpheme is attached to any type
of modifier phrase and generates nominalised phrases that are translatable as
‘manner/way of [(pro)noun]’ (47),
‘[adjective] manner/way’ (48)-(49), or ‘manner/way that [relative clause]’ (50).
The resulting manner
phrases can assume any syntactic function and any semantic role in the clause.
Ex. (47) contains two manner-nominalised phrases, the subject noun phrase Sabir-ó=g-u ‘Sabiro’s way’ and the
ablative-marked adverbial noun phrase Shaameeb-í=gíichch
‘from/than Shaameebo’s way’.
Apart
from genitive (pro)nouns, the manner nominaliser takes adjectival phrases as
input. Most frequently, =g is used
with the adjectives ‘good’, i.e. danáam-o=g(g)-a
‘well, in a good way’ (49), and ‘bad’, i.e. fárr-a=g(g)-a (48) and híil-a=g(g)-a
‘badly, in a bad way’. Apart from these three adjectives, =g is attested in 14 other adjectives in
the Kambaata schoolbooks (K89).
In
Kambaata, modifying adjectives always agree with their head noun (Treis 2008: 88-93) – and thus with the manner nominaliser – in case and
gender. The oblique case of the adjective in (48) and the nominative case of
the adjective in (49) are triggered by the accusative/oblique and the nominative
case markers of the enclitic =g,
respectively.[19] As seen in (47)-(49),
the manner morpheme is realised either as a single =g or a geminate =gg in
free variation. The geminate =gg is a
frequent variant of =g when the
stress falls on the penultimate syllable of the host. Finally, the manner nominaliser takes clausal inputs; more
precisely, it is added to modifying (relative) clauses. The nominalised entity
is used in object function in (50).
4.3.
The nominal origin of the manner nominaliser =g
Even
though =g is no longer used as an
independent word, it is very likely that =g
goes back to a fully-fledged noun ‘manner, way of doing something’. The
original form is, however, unknown. Sidaama, a closely related HEC language, has
a noun gara (m.) ‘manner, way of
doing something’ (Gasparini 1983: 114, Kawachi forthcoming), whose Kambaata
cognate may have served as the source for =g.
The nominal origin of =g is
reflected in its case-marking potential (Table 3). It can also host morphemes
that are found on regular nouns, e.g. the additive morpheme ‘also, too’ (ADD) (51),
the interrogative ‑’nnu-morpheme ‘and
what about?’, and the pragmatically determined -n-morpheme (N) (59).
The position of the copula in non-verbal clauses provides
further evidence for the nominal origin and status of =g. In (52), the manner nominalisation (in curly brackets) is the
predicate; the headless, nominative-marked relative clause is the subject. The
masculine gender allomorph of the non-locative copula -a (mCOP2) is triggered by the masculine gender of =g.
According to a general syntactic rule in Kambaata,
the non-locative copula (COP2) is found after derivational and inflectional
morphemes towards the end[20] of the predicate if the predicate consists of a single morphological
nominal word. It shifts to the preceding word if the predicate is modified by a
genitive noun, adjective, numeral or relative clause (Treis 2008: 414f). In (49), the copula is found predicate-medially, in
other words, on the relative modifier – which is evidence that the manner nominalisation is still considered a multi-word noun phrase
consisting of a modifier and a head noun rather than a single morphological
word. 4.4.
Predicative similative construction
Returning
to the similative construction, this section discusses how comparee and
standard of comparison are marked morphologically and which syntactic functions
they can assume. The focus here is on the predicative construction with a
phrasal standard of comparison (53). See §4.5 for the attributive construction
and §4.6 for the predicative construction with a clausal standard of
comparison. (53) Predicative similative construction
The
standard of comparison is an adverbial phrase to the predicate. The standard
marker =g is always followed by an
ACC/OBL case morpheme in the similative construction, not only in (54)-(55) but
in all the examples in my corpus. Whereas the syntactic function of the
standard of comparison is set, the comparee occurs in different syntactic
functions. In (54)-(55) the comparee is the subject of the clause and as such it
is marked by subject agreement on the verb; see 1s in (54) and 2s in (55). In
contrast, the comparee ha’mmichchús ‘enset
corm’ functions as the direct object in (56): ‘One boils the enset corm like
(one boils) potatoes’. But not: *‘One boils the enset corm like potatoes (boil
the enset corm)’.
In (57), the comparee is the unexpressed indirect
object of aass- ‘give’, namely the
children understood from the context. If they had been expressed overtly, they
would have been encoded in the dative case. In (58), two possessors, hair and
grass, are in a comparee-standard relation.
The comparee, plaaneet-áan
‘on the planet’, is a locative adverbial in (59).
The comparee can also be a temporal adverbial. Example
(60) does not compare the similarity in manner of two entities with respect to
a predicate but the similarity of a situation or an event at two different
points in time, i.e. this year and last year.
To summarise, the invariably ACC/OBL-marked
standard phrase gives no indication about which other explicit or implicit noun
phrase in the clause is the comparee for which it serves as the standard of
comparison. By definition, similative constructions express sameness of manner
(Haspelmath & Buchholz 1998: 278), whereby manner needs to be understood in
a broader sense, i.e. not merely as the techniques, the instruments and the
means that are applied and the type of movements (motor patterns) that are
carried out. For example, in (54), the compared entities share the same
instrument (a shard), and probably also the same technique (licking) and body
posture for drinking. In (56), the compared entities share the same means of
preparation, namely boiling water. Sameness of manner could also mean that the
actions are carried out at the same rate (57) or that the disposition, the
attitude or other psychological, social and physical conditions are shared by
the compared entities. One can, therefore, argue that the morpheme =g in its function as standard marker in
constructions expressing comparison of similarity, as illustrated in the
examples in this section, has undergone semantic extension. While =g is a nominaliser of manner in the
narrow sense of the word (techniques, instruments, means, types of movement) in
the non-comparison examples in §4.2, the manner encoded by =g in similative constructions embraces the broader sense of the
word. The comparee and the standard of comparison may, for instance, have
similar properties and possess the same attributes, e.g. in (58) and (59) and they
may have a similar shape, habits, needs, intentions, etc. A narrow manner
interpretation is often not possible in similative constructions. In (61), the respect
in which comparee and standard are similar is left completely open to interpretation.
They could, for instance, carry out an action in the same way, to the same
extent, with the same goal, or just share a certain physical, psychological, or
social quality.
4.5.
Attributive similative construction
All
elements of a similative construction may be found in one noun phrase (62). (62) Attributive similative construction
In
the attributive construction, ‘an X which V-s like a Y’, the comparee serves as
the head of the noun phrase, see beet-ú ‘son’
in (63), and the standard of comparison, marked by =g, is expressed in a relative clause modifying the head noun.
If a ‘be’ verb is required in the relative clause, ‘an X which is like a Y’, the locative copula yoo- ‘be located, exist’ (COP1) is used. It would be ungrammatical to drop the relative verb yóo in the attributive similative construction in (64), as a standard phrase cannot directly modify a noun (as in English ‘there are students like Awwano’).
If
an attributive similative construction is headless, it is nominalised with a
dummy head, e.g. =r ‘thing(s), ones’,
and thus serves to express ‘N-like thing(s), N-like one(s)’ (65).
The
use of yoo- (COP1) in the attributive
similative construction (64)-(65) is especially noteworthy because this copula
is otherwise found only in constructions expressing location, existence and
possession (Treis 2008: 398-407). Instead of yoo-, one would have expected the use of the verb ih- ‘be(come)’, which is the fully inflectable
verbal substitute of the ascriptive and identificational copulas COP2 and COP3
in subordinate clauses (Treis 2008: 427ff). 4.6.
Similative clauses
The
standard of comparison is not necessarily a (pro)noun phrase but can also be an
entire clause (66)-(67). Similative clauses are relativised clauses plus the
standard marker =g.
Similative clauses are used in a productive idiomatic sentence pattern which expresses that the subjects do something in a way the situation permits them to do it, which is always interpreted as an uncomfortable, unsatisfactory way of doing something (68). In these patterns, the subordinate clause contains a causative form of the verb in the syntactically superordinate clause. The subject of the subordinate clause is only marked by subject agreement on the verb (3m) but remains otherwise unexpressed; it can be assumed to be magán-u (mNOM) ‘God’. The subordinate verb is in the perfective form in all examples, while the aspect marking in the superordinate clause may vary.
5. Comparison of equality
By
comparison of equality I mean a type of comparison “that ascribes to the
comparee the same value of the
parameter of comparison as to the standard”, as stated by Cuzzolin &
Lehmann (2004: 1213, emphasis mine). In other words, following Henkelmann
(2006: 371), a quality is attributed to comparee and standard to an “equal
extent”. Kambaata distinguishes between two types of equative constructions (69).
In the first type, the standard of comparison is dependent on the enclitic
nominaliser =g ‘manner’ (§4), in the
second type, it is dependent on the noun qax-á
‘extent’. (69) Predicative equative constructions
The manner nominaliser =g has extended its function from a
marker of standards of similative comparison (§4) to a marker of standards of
equative comparison (69i). Equative and similative constructions differ only in
one respect: In equative constructions the parameter of comparison is overtly
expressed by adjectives, inchoative-stative property verbs or ideophones. The syntax of the two equative constructions in (69) is
identical. As in the case of the comparative construction (§3), the standard
phrase is an adjunct to the predicate expressing the parameter; inside the
standard phrase the standard marker is the head, while the standard is a
genitive modifier (possessor). The comparee is encoded as the subject; it
usually precedes the standard NP, but may be shifted
into the pre-predicate slot if focussed (cf. §3.1.1). Both constructions lack a
parameter marker. In Henkelmann’s typology of equative comparison (2006: 377), they
would thus belong to type I.A (comparee and standard in separate
NPs, parameter encoded as predicate, explicit EQUAL concept). In Haspelmath et
al.’s (2017) typology, the constructions discussed in this section fall under
Type 1 (“Only equative standard marker”), which is the most frequent
cross-linguistically. The constructions in (69) differ with regard to their frequency
as well as the origin, multifunctionality and degree of grammaticalisation of
their respective standard markers. These aspects will be treated in the
following sections. Both constructions are attested in spontaneous, written and
elicited data. 5.1.
Equative construction with =g
‘manner’
In the first type of equative construction, the
standard phrase is marked by the manner nominaliser =g, which is added to a
genitive-marked standard (pro)noun phrase (70)-(71).
The nominaliser itself is marked for the ACC/OBL-case and may
carry the pragmatically determined morpheme -n. In (70)-(71), the parameter is expressed by an adjective, and in
(72) by a property verb.
In (73), the adjectival parameter xalig-á ‘strong’ is the predicate of a
subordinate conditional clause and combines with a ‘be(come)’-verb, which
carries subject agreement and subordinating morphology. If the parameter was a
property verb or a property ideophone accompanied by a support verb, it could
itself carry the morphology required of predicates in subordinate clauses.
In
attributive equative constructions, all components of the comparison (comparee,
standard and parameter) are found inside one NP. In (74), the comparee billaww-á ‘knife’ is the head of the object NP of the clause. It is
modified by an adjectival parameter iphph-á
‘sharp’, which is preceded by the =g-marked
standard phrase in adverbial function.
Examples
in which the clausal standard of an equative construction is marked by =g are not attested in my database; see
instead §5.2.3. 5.2.
Equative construction with qax-á ‘extent’
In the second type of equative construction, the
standard phrase consists of the accusative-marked noun qax-á ‘extent’, which is preceded by a genitive modifier expressing
the standard of comparison; see Da’llis-ó
‘of Da’lliso’ (2sGEN) in (75) and isé
‘her’ (3fGEN) in (76).
Even though equative examples with qax-á have been readily provided by
Kambaata speakers in elicitation sessions, they are much less common than
equative construction with =g in
spontaneously produced data; note, however, that (75) is a non-elicited
example. It is not clear whether equality expressions with qax-á should be interpreted as a dedicated, conventionalised
construction. In particular, it would be interesting to explore the semantic
parameter type that qax-á can be used
with, and the type of equative relation it can express. In spite of elicited
examples such as (76), in which the parameter of comparison is colour, qax-á seems to be used primarily for the
expression of equality in size and equality of amount. Furthermore, if the
parameter of comparison is not overtly expressed (77)-(78) and if qax-á thus functions as the predicate of
the equative construction, the implicit parameter of comparison is always understood
to be size or amount, which shows that qax-á
is not (yet?) a semantically void standard marker.[22]
It is safe to say that the standard marker qax-á ‘extent’ is hardly grammaticalised. Unlike the manner nominaliser =g used in the first type of equative
construction, it is not phonologically eroded and shows no signs of being encliticised
to the standard. As we will see next, it still qualifies as a full noun. 5.2.1. The
morphology and polysemy of qax-á
Qax-á ‘extent’ is a masculine noun with full case-marking
potential (Table 4). As such, it can be used in any syntactic function in the
clause.
Table 4. Case paradigm of qax-á (m.) ‘extent’ The noun qax-á
is highly polysemous and is used with the concrete lexical meanings ‘amount,
quantity, number’ (79),[23] ‘volume,
contents’, ‘size, height, length, circumference’, ‘(spatial)
distance, limit’, and, as seen in (80), ‘(temporal) distance,
period of time, while’.
The noun qax-á is also used with a more abstract
meaning ‘degree’ in (81).
The noun qax-á can also be an adverbial modifier
of quality verbs, e.g. cúlu ass- ‘be
appealing (to somebody)’ in (82).
In accordance with
universal processes of semantic change, it seems reasonable to assume that qax-á originally had a fairly concrete
meaning of ‘amount’, ‘size’, ‘distance’, ‘period of time’ and only later
developed the more abstract sense of ‘degree’, which facilitates its use in the
expression of equality. 5.2.2. The equative
nomino-adjective qax-á(-ta)
Based on the stem of the noun qax-á illustrated in §5.2.1, Kambaata also has a dedicated equative
nomino-adjective qax-á(-ta)
‘which is/are (about) as much/many as, which has (about) the amount/size/ distance,
etc. of’.[24]
As a modifier, it agrees in case and gender with its head noun. In (83) the
oblique case of qáx-ata signals
agreement with a feminine non-nominative/non-accusative noun, more precisely
with ma’nnéen fLOC ‘at a place’.
Furthermore, it agrees in gender with its subject if it is used predicatively;
see yamáz-ut ‘waist’ (84), which
triggers the feminine predicative form of qax-á(-ta).
The nomino-adjective qax-á(-ta) is
often interpreted as expressing only approximate equality.
Unlike other adjectives in
Kambaata, which can govern accusative, dative and ablative complements (Treis
2008: 90), qax-á(-ta) combines,
unexpectedly, with a genitive phrase, as illustrated by meetir-í ‘of metre(s)’ in (83). Other than that, only nouns (or
enclitics of (pro)nominal origin) are modified by genitive phrases. Therefore,
the word class status of qax-á(-ta) cannot
be determined unequivocally and it is called a nomino-adjective. 5.2.3. Equative
clauses with qax-á ‘extent’
If
a clause headed by a finite verb expresses the standard of comparison in an equative
construction, it is relativised and the clause is thus
made the modifier of qax-á ‘extent’.
The standard marker occurs either in the accusative (85) or in the ICP case (86).
My
corpus contains only quantitative equatives of the type ‘V as much as’ with a
clausal standard where the parameter of comparison is implied in the standard
marker qax-á ‘extent’. Equatives with
a clausal standard and an explicit non-quantitative parameter, e.g. ‘She is as
intelligent/beautiful/nasty as we had assumed’, are not attested. Neither do equative
constructions with a clausal standard marked by =g ‘manner’ (§6.1) occur. Note, however, that =g cannot replace qax-á in the examples in this section. 6. Summary and cross-Cushitic comparison
The
Kambaata constructions used for the expression of the four types of comparison
studied in this paper, i.e. comparison of relative superiority, absolute
superiority (superlativity), equality and similarity, have been shown to have
standard markers that are grammaticalised to different degrees. In the comparative
construction, the standard phrase is marked by a true case morpheme, the
ablative case, which is one of the nine cases for which nouns in Kambaata
inflect. The same case morpheme is also used to mark the standard in one type of
superlative construction, which is based on the comparative construction and in
which the universal quantifier ‘all’ is the standard (‘X is big from all’). In
the second type of superlative construction, the standard of comparison is
marked by the locative case (‘X is big among Y’). Like Kambaata, many East Cushitic languages mark the
standard of comparison in comparative constructions by an ablative case
morpheme or an ablative adposition; see K’abeena (Crass 2005: 296), Alaaba (Schneider-Blum
2007: 94f), Hadiyya (Sim 1989: 354), Afar (Hassan Kamil 2015: 363), Dullay
(Amborn et al. 1980: 93, 101), Ts’amakko (Savà 2005: 127), and Somali (Saeed
1993: 107, 191f). The ablative is also one possible – but not the most common –
standard marker in Sidaama (Kawachi 2011: 97). Furthermore, standards are ablative-marked
in the Central Cushitic language Xamtanga (Darmon 2015: 324). In the East Cushitic
group we also find languages that make use of locative
adpositions as standard markers; see Konso (Ongaye 2013: 201, 179), Dhaasanac
(Tosco 2001: 293) and Oromo (Owens 1985: 28, 77, 128, 233, 265 and elsewhere on
Harar Oromo; Gragg 1982: 226 on Wellegga Oromo; Stroomer 1987: 50, 333 on Boraana
Oromo). These languages contest Zelealem & Heine’s (2003: 60) claim that “the
Source Schema provides the only or one of the main options for encoding
comparatives [in the Ethiopian area]”. The South Cushitic language Iraqw has a
dedicated comparative morpheme (Mous 1993: 214, 227). In Beja, the only North
Cushitic language, the origin of the standard marker, which is also used as a
parameter marker, is opaque. It may, however, be related to ablative morphemes
in related languages (Vanhove 2017). In the Cushitic
literature, information on superlative constructions is sparse. The
K’abeena and Alaaba grammars mention the use of comparative constructions with
a universal quantifier (lit. ‘X is big from all’) for the expression of superlativity
(Crass 2005: 297, Schneider-Blum 2007: 95). By way of contrast, in some
Cushitic languages, we also find constructions without a universal quantifier
and with unexpected structural differences between comparative and superlative
constructions. While the Somali comparative construction marks the standard
with an ablative preposition ká, the
superlative standard is marked by an unrelated prepositional cluster úgú, which might have originated from a
combination of a dative ‘to, for’ and a locative preposition ‘in(to), on, at,
with’ or a reduplicated dative preposition (Saeed 1993: 192, 201, 206). The
illustrative sentences in Gragg’s Oromo dictionary (1982) contain two kinds of
superlatives. Some are based on the regular comparative construction with irra ‘on’ and with a universal
quantifier as standard. Others use the locative postposition keessa ‘in(side), into’ > keessa-a (in-ABL) ‘from among’ (87).
Finally, the Iraqw grammar contains a superlative example
(Mous 1993: 214) with a locative preposition bará as a standard marker and a universal quantifier as part of the
standard phrase. This construction is not based on the regular comparative
construction with the dedicated standard marker ta. In the Kambaata similative construction, the standard marker
is not a case morpheme but a monosyllabic phrasal enclitic =g of nominal origin. The morpheme can no longer be used without a host
and has probably undergone phonological reduction in recent times (Kambaata
does not have any monosyllabic nouns); it has, however, retained most features
of a noun. The similative standard marker =g
is also the most common marker of the standard in equative constructions.
Alternatively, Kambaata speakers make use of an equative periphrasis of the
type ‘X is tall to the extent of Y’. The noun qax-á ‘extent’ does not seem to have undergone any phonological
reduction and is still used as a full, albeit semantically polysemous (or
vague) full noun. Treis (2017b)
shows that in Central Cushitic and Highland East
Cushitic, the standard of comparison in similative constructions is marked by
postposed morphemes, whereas in Lowland East Cushitic languages (with the exception
of Afar and Saho) it is marked by preposed morphemes. Some Cushitic languages
have two different similative morphemes that are used interchangeably in certain
contexts (see also Kawachi forthcoming on Sidaama). The grammatical status of
the standard marker varies from language to language. Whereas it is a
(semi-)dependent morpheme in most Cushitic languages, in other languages it has
a primary use as a free noun meaning ‘manner’. And even if a language cannot
(or can no longer) use its similative morpheme as a full noun, its presumed
nominal origin may still be reflected in its case-marking potential or the
formal marking of the standard that it governs (which is often marked for the
genitive case). In 12 out of 20 Cushitic languages investigated by Treis (2017b)
the similative standard marker is also employed as standard marker in equative
constructions. Alternatively, some Cushitic languages use a noun ‘extent’ (or
similar) as the head of the equative standard phrase; see Vanhove (2017) for Beja. This paper has only been able to touch on some of the
functions of the Kambaata enclitic morpheme =g
‘manner; like’ in the section on similative and equative comparison. The
morpheme is, however, highly multifunctional; a detailed description of this
multifunctionality is provided in Treis (2017b). In the same publication some
preliminary information can also found on the expression of hypothetical
similarity (‘do as if’) (Treis 2017b: §2.4). A more thorough investigation of
this sub-type of comparison of similarity is still required. Abbreviations
References
Amborn, Hermann, Gunter Minker & Hans-Jürgen Sasse
1980. Das Dullay. Materialien zu einer ostkuschitischen
Sprachgruppe. Berlin: Reimer.
Crass, Joachim 2005. Das
K’abeena. Deskriptive Grammatik einer hochlandostkuschitischen Sprache. (Kuschitische Sprachstudien, 23.) Cologne:
Köppe. Chelliah, Shobhana L. & Willem J. de Reuse 2011. Handbook of Descriptive Linguistic Fieldwork.
Dordrecht et al.: Springer. Cuzzolin, Pierluigi & Christian Lehmann 2004.
Comparison and gradation. In: Booij, Gert et al. (eds.), Morphologie. Ein internationales Handbuch zur Flexion und Wortbildung,
vol. 2, pp. 1857-1882. Berlin, New York: W. de Gruyter. Darmon, Chloé 2015. A Morphosyntactic Description of Xamtanga, an Agaw (Central Cushitic) Language
of the Northern Ethiopian Highlands. PhD thesis. Lyon: Université Lumière
Lyon 2. Dixon,
R. M. W. 2008. Comparative constructions: a cross-linguistic typology. Studies in Language 32, 4: 787-817. Fuchs, Catherine 2014. La comparaison et son expression en français. Paris: Ophrys. Gasparini, Armido 1983. Sidamo-English Dictionary. Bologna: EMI. Geetaahun Heelleebo Baachchoore 2002 E.C. Kambaatissa Afgale Maxaafa [Collection
of Kambaata Oral Literature]. Tungu: Kambaatina
Xambaaro Zoonaan, Dooyyoganni Woradaan, Maare-Ammachcho Waato. Gorshenin, Maksym 2012. The crosslinguistics of the
superlative. In: Stroh, Cornelia (ed.). Neues aus der Bremer Linguistikwerkstatt: Aktuelle Themen und Projekte
31, pp. 55-160. Bochum: Brockmeyer. Gragg,
Gene B. (ed.) 1982. Oromo Dictionary.
East Lansing, MI: African Studies Center, Michigan State University. Haspelmath,
Martin & Oda Buchholz 1998. Equative and similative constructions in the
languages of Europe. In: van der Auwera, Johan (ed.). Adverbial constructions in the languages of Europe, pp. 277-334.
Berlin: Mouton de Gruyter. Haspelmath,
Martin & the Leipzig Equative Constructions Team 2017. Equative constructions
in world-wide perspective. In: Treis, Yvonne & Martine Vanhove (eds.). Similative and Equative Constructions: A
Cross-linguistic Perspective, pp. 9-32. Amsterdam, Philadelphia: Benjamins.
Hassan Kamil, Mohamed 2015. L’afar: Description
grammaticale d’une langue couchitique (Djibouti, Erythrée et Éthiopie). PhD thesis. Paris:
INALCO. Heine,
Bernd 1997. Cognitive Foundations of
Grammar. New York, Oxford: Oxford University Press. Henkelmann,
Peter 2006. Constructions of equative comparison. Sprachtypologie und Universalienforschung 59, 4:
370-398. K89 = Kambaatissata.
Rosaanchi Maxaafa. [Kambaata Language. School Book.] 1989 E.C. Grade
1-8. Southern Nations, Nationalities, and Peoples Regional State: Education
Bureau. Kambaata
and Hadiyya Translation Project Hosaina 2005. Yohaannis – Latin Version of
the Gospel of John in Kambaata Language. Addis Ababa: The Bible Society of
Ethiopia. Kawachi, Kazuhiro 2007. A Grammar of Sidaama (Sidamo), a Cushitic Language of Ethiopia. PhD
thesis, University at Buffalo, the State University of New York. Kawachi, Kazuhiro 2011. Can Ethiopian languages be
considered languages in the African linguistic area? In: Hieda, Osamu, Christa
König and Hirosi Nakagawa (eds.). Geographical
Typology and Linguistic Areas – with Special Reference to Africa, pp. 91-107.
Amsterdam, Philadelphia: John Benjamins. Kawachi, Kazuhiro forthcoming. Sidaama. In: Tsunoda,
Tasaku (ed.). Mermaid Constructions.
(Comparative Handbooks of Linguistics) Berlin: De Gruyter Mouton. Mous, Maarten 1993. A Grammar of Iraqw. Hamburg: Buske. Ongaye Oda Orkaydo 2013. A Grammar of Konso. Utrecht: LOT. Owens, Jonathan 1985. A Grammar of Harar Oromo (Northeastern Ethiopia). Hamburg: Buske. Saeed, John 1993. Somali
Reference Grammar. Kensington: Dunwoody Press. Saint-Exupéry, Antoine de. 2018. Qakkichchu Laaha [The Little Prince]. Translated by Deginet Wotango
Doyiso & Yvonne Treis. Neckarsteinach: Tintenfaß. Savà, Graziano 2005. A
grammar of Ts’ amakko. (Kuschitische Sprachstudien, 22.) Cologne: Köppe. Schneider-Blum, Gertrud 2007. A Grammar of Alaaba, a Highland
East Cushitic Language of Ethiopia. (Kuschitische
Sprachstudien, 25.) Cologne: Köppe. Sim, Ronald J. 1989. Predicate
Conjoining in Hadiyya: A Head-Driven PS Grammar. PhD thesis. Edinburgh: University of
Edinburgh. Stassen, Leon 1985. Comparison and Universal Grammar. Oxford: Basil Blackwell. Stassen,
Leon. 2013. Comparative Constructions. In: Dryer, Matthew S. & Martin
Haspelmath (eds.). The World Atlas of
Language Structures Online. Leipzig: Max Planck Institute for Evolutionary
Anthropology. Available online at http://wals.info/chapter/121, accessed on
2015-01-07. Stroomer,
Harry 1987. A Comparative Study of Three
Southern Oromo dialects in Kenya: Phonology, Morphology and Vocabulary.
Hamburg: Buske. Tosco, Mauro 2001. The
Dhaasanac Language. (Kuschitische Sprachstudien, 17.) Cologne: Köppe. Treis,
Yvonne 2008. A Grammar of Kambaata. Part
1: Phonology, Morphology, and Non-verbal Predication. (Kuschitische
Sprachstudien, 26.) Cologne: Köppe. Treis, Yvonne 2012. Categorial hybrids in Kambaata. Journal of African Languages and Linguistics 33, 2: 215-254. Treis,
Yvonne 2017a. “They are only two, like the teats of a donkey”: Kambaata
denumerals revisited. In: Kramer, Raija & Roland Kießling (eds.). Mechthildian approaches to Afrikanistik:
Advances in language based research on Africa. Festschrift in honor of Mechthild Reh, pp. 339-366.
Cologne:
Köppe. Treis,
Yvonne 2017b. Similative morphemes as purpose clause markers in Ethiopia and
beyond. In: Treis, Yvonne & Martine Vanhove (eds.). Similative and Equative Constructions: A Cross-linguistic Perspective,
pp. 91-142. Amsterdam, Philadelphia: Benjamins. Vanhove, Martin 2017. Similative, equative, and
comparative constructions in Beja (North-Cushitic). In: Treis, Yvonne & Martine
Vanhove (eds.). Similative and Equative
Constructions: A Cross-linguistic Perspective, pp. 189-211. Amsterdam,
Philadelphia: Benjamins. Wedekind, Klaus 1990. Generating Narratives: Interrelations of Knowledge, Text Variants and
Cushitic Focus Strategies. Berlin: Mouton de Gruyter. Zelealem Leyew & Bernd Heine 2003. Comparative
constructions in Africa: An areal dimension. Annual Publication in African Linguistics 1: 49-68. [1] Financial support for this work came from the federation Typologie et universaux linguistiques:
données et modèles (CNRS, FR
2559) via the project Expression des
comparaisons d’égalité et de similitude (2014-2018). I am grateful to the Culture
Department of the Kambaata-Xambaaro Zone for their support during my
fieldtrips. Meaza Kerlos collected most of the written Kambaata sources. Mirja
Saksa obtained the Kambaata translation of the Gospel of John for me. I am indebted to my
language assistant, Deginet Wotango, for the intensive and fruitful discussions
of the analyses proposed here. I would like to thank Teshome Dagne, Tessema
Handiso and all the other Kambaata speakers I have been working with since
2002. Claudine Chamoreau, Kazuhiro Kawachi, Tatiana Nikitina, Martine Vanhove
and Kasia Wojtylak gave valuable feedback on an earlier version of this paper. [2] The sources of my corpus data are given in [square
brackets]. [3] Local publications: K89 = Kambaatissata (1989), Geetaahun
(2002), Saint-Exupéry (2018), and John = Gospel of John (Kambaata and Hadiyya
Translation Project Hosaina 2005). [4] The terms “property ideophone” and “property verb” include
those members of the word classes of ideophones and verbs that are cognate with
(i.e. have the same stem as) an adjective, e.g. gambáll y- (ideophone) ‘be(come) black’ – gamball-á(-ta) (adjective) ‘black’; qeraa’rr- (verb) ‘be(come) tall, long’ – qeraa’rr-ú(-ta) ‘tall, long’.
[5] According to a native speaker,
abb-á ‘big; much’ is not obligatory
in (8). It implies here that both men are rich, while this implication is
absent without abb-á (in this case
one of the two would be expected to be poor). [6] As suggested by an anonymous
reviewer, one could also assume that the standard is topicalised in (9). The
pragmatic factors determining the constituent order remain to be explored. [7] The verb abbis- ‘exceed’ is the causative form of abb- ‘be(come) big, great, honoured’. [8] Baas-á can also be used as
an intensifier in adverbial function (9). [9] Inherently comparative lexemes
meaning ‘(be/become) better’, some of which are cognate to the Kambaata lexeme,
can be found in related East Cushitic languages, e.g. K’abeena/Alaaba k’oh- ‘be better’ (Crass 2005: 296,
Schneider-Blum 2007: 95), Sidaama/Gedeo woyya
‘better’ (Kawachi 2007: 38, Wedekind 1990: 405), Konso woyy- ‘be better’ (Ongaye 2013: 167), Oromo wayya ‘be better’ (Gragg 1982: 402). [10] Aluud-ú ‘above; north’ and alúud-iin ‘above’ are used outside the
context of comparison as locative adverbials; they have resulted in the merger
of al-í wud-ú (up-mGEN side-mACC) and
al-í wud-íin (up-mGEN side-mICP). [11] According to a native speaker, the converb hígg ‘passing’ could be dropped without
a change in meaning. [12] Including verbs accompanying
ideophonic parameters. [13] Treis (2012) provides more
information on negative relative verbs (participles). [14] I have chosen ‘among X’ as the
literal translation of the locative-marked standard. Given that the Kambaata
locative expresses a vague locative relation, I could have also chosen ‘in X’,
‘on X’ or ‘at X’. [15] The -n-morpheme is frequently found on nouns and pronouns of all cases
except the genitive, on modifying and predicative adjectives and numerals, and
on subordinate verb forms. There are even instances where it is attested on
main verbs (for a preliminary analysis see Treis 2008: 220-27). [16] Treis (2008: 285ff, 379f, 388f) provides information
on the exceptional cases where the manner morpheme is added to a demonstrative
root and where it thus has characteristics of a derivational morpheme rather
than a phrasal enclitic. [17] See Treis (2008: 103) for an overview of the nine
feminine and twelve masculine nominal declensions. [18] The ACC and the OBL forms cannot be differentiated if
the case marker is the last morpheme of the word. However, the presence of
additional suffixes, e.g. the pragmatically determined morpheme -n, causes a distinction to surface again
between the ACC and OBL forms, which was presumably made in an earlier stage of
the language. The combination of =g-a
OBL and -n results in =g-a-n (67). In contrast, in the ACC
case, the -n is infixed into the
older, non-eroded ACC form =g-aha,
which results in =g-anka (59). [19] Cardinal numerals and
demonstratives, which are also case/gender-agreeing modifiers, cannot be the
input of the manner nominalisation. Native speakers reject as ungrammatical
forms such as *lám-u=gg-u
two-mNOM=G-mNOM ‘two ways’ or *ku=g-u
A_DEM1.mNOM=G-mNOM ‘this way’. It is, however, possible to add the manner
nominaliser to ordinal numerals, e.g. woná=g-u
‘the first way’ (NOM), lankí=g-u ‘the
second way’ (NOM) etc. Ordinal numerals do not show case/gender-agreement; they
share features of genitive nouns (for details see Treis 2008: 236f, 313ff;
2017a). [20] The copula can still be followed by possessive,
coordinative and pragmatically determined suffixes. [21] A native speaker noted that Da’llis-ó qax-á could here be replaced by Da’llisó=g-a (§5.1) without any apparent change in meaning. [22] Note that size/amount polysemy is also observed in the
adjectives qah-ú(-ta) ‘small (size);
little (amount)’ and abb-á(-ta) ‘big;
much’. [23] For a discussion of the use of qax-á to express approximate number see Treis (2017a: §9). [24] Cf. dedicated quantitative equatives in European
languages, as discussed in Haspelmath & Buchholz (1998: 298-301). | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||