Issues in Dagbanli Compounding

In studying compounds, complex words formed from already existing words, many interesting issues come up, touching virtually every aspect of the study of language—phonology, morphology, syntax, semantics, and sometimes, pragmatics. These issues may relate to the characterization of compounding, their classification, meaning, whether or not they are headed constructions as well as how to distinguish them from derived words on the one hand and phrases on the other. This study discusses some core issues in studying Dagbanli compounding, a Mabia language spoken in the Northern Region of Ghana. It deals with issues such as the classification of Dagbanli compounds, headedness in Dagbanli compounds: types and position and distinction between compounds and phrases in Dagbanli. Regarding the possible classes of compounds in Dagbanli, it is shown that the most common compounds consist of only two lexical bases and that the commonest combinations of syntactic categories in Dagbanli compounds are Noun-Noun, Noun-Adjective, Noun-Verb and Verb-Noun. It is also shown that while most compounds are right-headed (RH), phrases tend to have an obligatory head initial occurrence, making them left-headed (LH) constructions.

Introduction

This paper examines some core issues in the study of Dagbanli compounding. The language has an SVO word order and most of the inflectional affixes occur as suffixes (Hudu, 2014; Olawsky, 2002). The paper focuses on the issues on typology of Dagbanli compounds, headedness in Dagbanli compounds and the contrast between compounds and phrases in Dagbanli.

Dagbanli is a mabia language spoken by about one million, two hundred and fifty-eight thousand, six hundred and fifty-seven (1,258,657) native speakers (Inusahet al., 2020). The dialects of Dagbanli are Tomosili, Nayahili, and Nanunli. Inusah (2021) describes Gbanjonsili, Zundusili and Jimansili as its sub-dialects due to alternative pronunciations and grammar. The syllable types identified in the language are V, CV, CVC, CVV syllable. Syllabic nasals also occur in the language usually as [m], [n] or [ŋ], which are restricted to either pronominal or focus marking (Inusah, 2021, 2024; Mahama & Inusah, 2023).

In terms of the type of morphology, Dagbanli is morphologically an agglutinative language with some degree of fusion in which certain combinations of morphemes may be obscured by phonological and morphological rules typically affecting nouns and adjectives. Hudu observes: Dagbanli nouns and adjectives have the same morphology. In its simplest form, a noun/adjective consists of a root and a suffix that marks singular or plural number. The lexical meaning is provided by the root, although the suffix may serve to disambiguate homophonous roots (Hudu, 2014, p. 6).

Regarding the syntax, the observed structure typical declarative clause is an SVO word order and most of the inflectional affixes in the language occur as suffixes. In the next subsection, the phonological word in Dagbanli is discussed because it will be needed in the explication of the structure of Dagbanli compounds in subsequent sections.

The Dagbanli Word

Traditionally, Inusah (2016, p. 76) attests, “Dagbanli words can be classified into three major forms: simplex, complex and compound words. The simplex word consists of only one lexical root, which are mostly short in a CV syllable form without a coda or CVC syllable root with a coda, which does not require affix” as in Table I. “Nouns and adjectives in simplex forms consist of a lexical root and a number suffix bound to each other and as free-standing unit, the base is inflected with a singular or plural nominal suffix” (Hudu, 2014, p. 6) as illustrated in Tables II and III:

Table I. Verbs (Inusah, 2016, p. 76)

CV	Gloss	CVC	Gloss
tì	‘give’	tɨ́m	‘send’
dì	‘eat’	dɨ́m	‘bite’
tú	‘insult’	tɨ́b	‘heal’
dú	‘climb’	kàb	‘break’
té	‘filter’	màl	‘make’
zù	‘steal’	sàb	‘write’
tò	‘pound’	záŋ	‘take’
lù	‘fall’	láb	‘throw’

Table II. Nouns (cf. Inusah, 2016, p. 76)

Singular	Plural	Gloss
tú-á	tʊ́hí	‘baobabs’
ʃí-á	ʃɨ́hí	‘bees’
mì-á	mìhí	‘ropes’
tí-á	tíhí	‘trees’
bí-á	bíhí	‘children’
sú-á	sʊ́hí	‘knives’

Table III. Adjectives

Singular	Plural	Gloss
vìƐ-lí	vìƐl-á	‘nice’
jì-á	jí-hí	‘short’
bíl-á	bí-hí	‘small’
pál-ó	pàl-à	‘new’
sàb-lí	sàbɨ́l-á	‘black’
kʊ́ɾ-lí	kʊ́ɾá	‘old’

Theoretical Preliminaries

Aikhenvald (2007, p. 24) observes, “compounding involves word-formation based on the combination of at least two potentially free forms, most frequently members of open lexical classes such as nouns or verbs, e.g., English fox-hunting, station master.” Booij (2007, p. 75) characterizes compounding as “combination of two words, in which one word modifies the meaning of the other, the head.” Guevara and Scalise (2009) note, “compounds are the morphological constructions which are closest to syntactic constructions to the point that it is sometimes difficult to distinguish between phrases and compounds.” Consider the following examples illustrated in Table IV (Spencer, 1991, p. 310).

Table IV. Endocentric Compounds in Dagbanli

Taxi driver	→	Driver of a taxi
Poet painter	→	Poet and painter
Hard ball	→	A ball which is hard

In compound formation, nominal compounding results in the creation of new nouns while verb compounding forms new verbs and other word classes such as adjectives. Aikhenvald (2007) observes that compound formation is dominant in isolating languages, which typically has a one-to-one correspondence between morphemes and words.

Lieber (1992) mentions that endocentric and exocentric are identified as the two principal types of compounds. Dressler (2006) remarks that “exocentric compounds are marked relative to endocentric compounds and they are less common cross-linguistically and less common in the word classes in which they occur.” The data below illustrates endocentric compound which has a head element ʒɨ́m ‘blood’ in it as in Table Va and Table Vb is exocentric compound at the semantic level which does not have a head element in it in Dagbanli.

Table V. Endocentric and Exocentric Compounds in Dagbanli

Input	Output
a. nàhʊ́ ‘cow’ + ʒɨm ‘blood’	nàh-ʒɨ́m ‘cow blood’	from cow and blood
b. káɾɨ́m ‘read’ + bì-á ‘child’	kárɨ́m-bì-á ‘pupil’	from read and child

Syntactically, in Table Vb, the verb kárɨ́m ‘read’ is the head of the compound because it selects the noun bì-á ‘child’ as its complement; morphologically, the noun bì-á ‘child’ is the head since it is the locus inflection, this can be described as [[kàɾɨ́m-bí-hí ‘pupils’]] and not ^*[[kárɨ́m-hí-bì-á]]. The examples in Table VI show more endocentric compounds in Dagbanli where the final stem is the head element; the final stem of the endocentric compounds in Dagbanli functions as the morphological head, from which the plural is formed as presented below.

Table VI. Endocentric Compounds with Morphological Head in Dagbanli

a.	bàɣísí	kʊ́ɣ[ɨ́]-lí	→	bàɣísí-kʊ́ɣá
	grind	stone		grind	stones
	‘grinding stone’			‘grinding stone’
b.	kàɾɨ́m	bì-á	→	kàɾɨ́m-bí-hí
	read	child		read-children
	‘pupil’			‘pupil’
c.	vʊ́hɨ́m	dàbɨ́sɨ́-lí	→	vʊ́hɨ́m	dàb-sá
	rest	days		rest	days
	‘rest days’			‘rest days’

Previous studies on compounding have observed that the main issues in the study of compounding that have attracted the most scholarly attention and engendered lively debates are the typology, headedness and the distinction between compounds and phrases. The objective of this paper is to discuss this core issues in the study of compounding in Dagbanli.

Method

This paper explored the sociolinguistic interviews as data collection procedure as described by (Bauer, 2011; Hudson, 2001; Milroy & Gordon, 2003). This allowed the researcher to be part of the speech community. To elicit the data, the respondents were engaged in a natural conversation to collect the initial data while the sociolinguistic interviews were used on one-on-one exchanges between the researchers and the 20 participants who were selected for the interview to cross-check the initial data obtained. The interviews were done in natural settings such as festivals, funerals outdooring, weddings etc., where the participants were more comfortable. Each interview lasted not less than 30 minutes and the whole fieldwork lasted close to a month. The purpose of the interview was explained to each participant before each interview section. The data were collected in Yendi for the primary (solicited through the interviews) and secondary taken from other works (Hudu, 2014; Inusah, 2016, 2017, 2021, 2024; Olawsky, 1999, 2002). A total of 40 compound words in Dagbanli were collected for this paper and every analysis here is based on the data available.

Core Issues in Dagbanli Compounding

There are issues in the study of compounding of a language and for that matter that of Dagbanli where much is not studied in compounding. These issues in the study of compounding have drawn the attention of some scholars who have engaged in addressing them with diverse opinions. This paper discusses some core issues of compounding in Dagbanli focusing on typology, headedness and distinction between compounds and phrases in Dagbanli.

Typology of Dagbanli Compounds

Booij (2007) describes the process as a ‘combination of lexemes into larger words,’ which means a compound may comprise two lexical items put together in which one word modifies the head. This definition of a compound dwells on two major issues, that is, joining two words together and creating a new word with an independent or associated meaning from its constituent elements. In this case, compounding can be characterized as a combination of different roots or words to create a new word in which the meaning of the newly formed item may not necessarily reflect the meaning of its single components.

A free-standing word in Dagbanli has a number suffix marker attached to it, which is not the same in other languages like English. For instance, woman in English may be free standing but in Dagbanli the noun páɣá ‘woman’ needs a number marker to show its simplex form. So, the difficulty is that English may easily form a compound with woman as a constituent and it will still remain as a free form but in Dagbanli, the story is different. Donalies (2004) outlines definitive criteria of compound hood by providing a putative list of defining properties of compounds. A cursory look at the properties and relating property such as “a compound is formed without word-formation affixes” to compounding in Dagbanli, it is not clear what kinds of affixes are said to be absent in compound formation in the language since all forms of words either simple or complex are suffixed with number markers. In compounding in Dagbanli, final vowels in the first constituent of a compound is elided because of its location as presented below:

In Table VII, the constituents in the compounds in both (a) kɔ́ŋ-á + bì-á and (b) zɔ́ŋ-á + nàpɔ́ŋ-á each has a number marker –a to show their singular forms but the singular marker –a is deleted from the final position of the first constituent of each compound; the suffix marker in the second constituent is maintained to show the singular form of the compounds. Example in (Table VIId) is not accepted because the final vowel in the first noun must be deleted. It is shown in (Table VIIe) that compounds in Dagbanli cannot be formed without word-formation suffixes. There are other proves to show that affixed words can be constituents of compounds or that compounds can undergo further affixation. For example, Dagbanli compound as seen below bears the affix -ba. Besides, the base that -ba attaches to is itself a compound. Thus, the notion that “compound is formed without word-formation affixes” fails completely, and finding counter examples to the rest will be pretty easy in Dagbanli.

Table VII. Compounds with Word-Formation Affixes

	Input	Output
a.	kɔ́ŋ-á ‘leper’ + bì-á ‘child’	kɔ́ŋ-bì-á ‘leper’s child’
b.	zɔ́ŋ-á ‘bat’ + nàpɔ́ŋ-á ‘foot’	zɔ́ŋ-nàpɔ́ŋ-á ‘bat’s foot’
c.	kɔ́ŋ-á ‘leper’ + bí-hí ‘children’	kɔ́ŋ-bí-hí ‘leper’s children’
d.	kɔ́ŋ-á ‘leper’ + bí-á ‘child’	*kɔ́ŋ-á-bì-á ‘leper’s child’
e.	kɔ́ŋ-á ‘leper’ + bì-á ‘child’	*kɔ́ŋ-bí ‘leper’s child’

Following the data in Table VIII, it shows that the issue of defining compounding may be language specific since the internal structure of the languages are not the same though they might be some level of similarities. Compounds in Dagbanli as noted consist of only two lexical roots forms in which the final vowel of the first lexical item is reduced to a weak vowel or elided completely, this is shown in Table IX.

Table VIII. Compounds with Further Affixation

tʊ́m-tʊ́mdí-bá	làhɨ́n-ցʊ́
work-work-do-persons	group
‘trade union’

Table IX. Deleting the Final Vowel in the First Lexical Item in Compounds

Input	Output
nà-á ‘chief’ + bí-hí ‘children’	ná-bíhí ‘chief’s children’
nàhʊ́ ‘cow’ + zʊ́ɣʊ́ ‘head’	nàh-zʊ́ɣʊ́ ‘cow’s head’
wɔ́b-ցʊ́ ‘elephant’ + mà ‘mother’	wɔ́b-mà ‘elephant’s mother’
bá-á ‘dog’ + bɨ́n-dɨ́ ‘feaces’	bá-bɨ́ndɨ́ ‘dog’s feces’
báɣ-á ‘soothsayer’ + kɔ́l-ցʊ́ ‘sack’	báɣ-kɔ́l-ցʊ́ ‘soothsayer’s

As seen in Table IX, the final vowel segment in the first root of each compound is deleted but the second root form is not affected by any form of deletion.

Based on the syntactic category of the constituents of a compound, the possible structure of compounds is classified into three: noun–noun (N-N), noun–adjective (N-A) and verb–noun (V-N) which form nominals as argued to occur in Dagbanli by Olawsky (2002). In addition to the three compounds, the present data also identify noun–verb (N-V) compound. It is important to find answers to the obvious questions in compounding: which lexical categories are typically combined with each other and secondly, in which order are the categories combined? The possible combinations of words in compounds are limited and all compounds in Dagbanli are nouns, composed of nominal, adjectival, or verbal roots plus a morphological head. The head is always at the right and the plural formation of each word (where applicable), is realised on the head.

Table X shows the possible combinations of compounding in Dagbanli; it reveals that the attested types of compounds that exist in the language are N-N (noun-noun), N-A (noun-adjective), N-V (noun-verb) and V-N (verb-noun) compounds. N-N compounds consist of two nominal roots, the first element occurs without its class affix, whereas the second part keeps its regular suffix, being the morphological head. In most cases, the second part is also the semantic head, though some counter examples can be found.

Table X. Types of Compounds in Dagbanli

	N	A	V
N	N-N	N-A	N-V
A	*	*	*
V	V-N	*	*

N-N compounds seen in Table XI consist of a stem noun without a suffix marker –a, -u or -ցa and a suffixed stem with the suffix markers which mark for number to indicate that the compounds are realized in the singular form. These compounds can be inflected to create their plural forms by attaching the appropriate plural marker (e.g., –a, -hi etc.). For instance, the compounds: páʔ-ʧɨ́n-ʧɨ́nɨ ‘woman cloth’, nó-gán-lí ‘fowl egg’ when they are inflected with the plural marker –a, they are realized as ʧɨ́n-ʧɨ́n-a ‘women clothes and nó-ցál-á ‘fowl-eggs’.

Table XI. N-N Compounds in Dagbanli

Noun	Noun	Compound (N+N)	Gloss
páɣá ‘woman’	ʧɨ́n-ʧɨ́nɨ ‘cloth’	páɣ-ʧɨ́n-ʧɨ́nɨ́	‘woman cloth’
nóó-ցá ‘bird’	tìƐɣʊ́ ‘nest’	nóó-tìƐɣʊ́	‘bed nest’
nàh-ʊ́ ‘cow’	zʊ́ɣ-ʊ́ ‘head’	nàh- zʊ́ɣʊ́	‘cow head’
dà-á ‘market’	bì-á ‘child’	dà-bì-á	‘business person’
bàɣ-á ‘soothsayer’	kɔ́lցʊ́ ‘sack’	bàɣ-kɔ́l-ցʊ́	‘soothsayer sack’
zʊ́ɣ-ʊ́ ‘head’	pál-ó ‘plot’	zʊ́ɣ-pál-ó	‘bald head’
nó-ó ‘fowl’	ցàn-lí ‘egg’	nò-ցán-lí	‘fowl egg’
wàh-ʊ́ ‘horse’	dɨ́b-ցá ‘chewing stick’	wàɾ-dɨ́b-ցá	‘stallion’

An emerging issue suggests that in Dagbanli, N-N compounds and V-N compounds have comparable compound structure but not the same. For example, the word for ‘hail’, sà-kʊ́ɣ-lí, contains the elements sà-à ‘rain’ and kʊ́ɣ-lí ‘stone’. Similarly, verbs can form the initial part of a compound, as in bàɣ-sɨ́-kʊ́ɣ-lí ‘grinding stone’ where bàɣ-sɨ́ means ‘grind’. This claim may be addressed in future study with more data in Dagbanli compounds.

A productive type of compound in Dagbanli involves a noun root followed by an adjective. Plural or singular forms are marked only on the morphological head as seen in the data in Table III above. For instance, the singular of pál-lí ‘road’ ends with a singular marker -li and, therefore, the word is inflected according to number class -a, i.e., the plural will be pàl-á ‘roads. In many compounds, the second word has a number suffix maker.

The data in Table XII show the examples of N-A compounds which comprise a noun (e.g., dɔ́ɾó ‘disease’) and an adjective (e.g., kʊ́ɾ-lí ‘old’). The process of the compound formation is similar to that of the N-N in which the number marker is deleted. This indicates that N-A compounds interpretation would be a straight forward matter because they are each made up of a noun head which is modified by an adjective expressing some salient property of the referent of the noun. The data show that for a significant number of Dagbanli N-A compounds, Dagbanli does not have A-N compounds as ^*vɨ́Ɛl-páɣ-á ‘woman nice’ or ^*vɔ-li ɲe ‘nose hole’, which are ungrammatical.

Table XII. N-A Compounds in Dagbanli

Noun	Adjective	Compound (N+A)	Gloss
lʊ́ŋ-á ‘drum’	bɨ́lá ‘small’	lɔ́ŋ-bɨ́l-à	‘small drum’
páɣá ‘woman’	vɨ́Ɛlí ‘nice’	pàɣ-vɨ́Ɛlí	‘nice woman’
nó-ó ‘fowl’	pɨ́Ɛlí ‘white’	nò-pɨ́Ɛ-lí	‘white fowl’
bɔ́ɣʊ́ ‘arm’	bɨ́lá ‘small’	bɔ́ɣ-bɨ́lá	‘small arm’
d-ó ‘man;	bɨ́Ɛɣʊ́ ‘ugly’	dò-bɨ́Ɛ-ɣʊ́	‘ugly man’
dóɾó ‘disease’	kʊ́ɾ-lí ‘old’	dɔ́ɾ-kʊ́ɾ-lí	‘old disease’
ɲé é ‘nose’	vɔ́lí ‘hole’	ɲè-vɔ́-lí	‘nostril’

The N-V type of compound involves nominal and verbal base(s) as illustrated in Table XIII below in which the noun is attached to a verb to form the compound. This type displays certain productivity in that the verbal root is suffixed with a noun suffix of one of the inflectional classes. Plural (where applicable) is formed according to the respective class as presented in Table XIII.

Table XIII. N-V Compounds in Dagbanli

Noun	Verb	Compound (N+V)	Gloss
bɔ́ɣʊ́ ‘arm’	kábɨ́ ‘break’	bɔ́ɣ-kablɨ́	‘broken arm’
làɣɨ́ɾɨ́ ‘money’	dɨ́ɾa ‘chopping’	làɣ-dɨ́ɾá	‘chop money’
bì-á ‘child’	jʊ́ɾɨցʊ́ ‘loving’	bɨ́-jʊ́ɾɨ́ցʊ́	‘favourite child’
bɨ́nɨ́ ‘thing’	dɨ́ɾá ‘eating’	bɨ́n-dɨ́ɾá	‘foods’
tɨ́ŋa ‘floor’	vʊ́ɾɨ́ցʊ́ ‘crowler’	tɨ́n-vʊ́ɾɨ́ցʊ́	‘earth crowler’
sɔ́lɨ́ ‘street’	ʧándá ‘going’	sɔ́-ʧándá	‘traveller’
ʃè-è ‘waist’	lɔ́ɾɨ́ցɨ́ ‘tie’	ʃé-lɔ́ɾɨ́ցá	‘belt’

The semantic structure of the words in Table XIII is not homogeneous as some examples like bɔ́ɣ-káb-lɨ́ ‘broken arm’ resemble N-A compounds. Usually, the ‘nouns’ resulting from a combination of verbal root plus derivational suffix cannot occur without the first lexical root in a compound. Also, in the N-V compounds, suffix markers are only attached to the verb.

V-N is another type of compounding that is very rare in Dagbanli. Olawsky (1999) suggests that because the V-N compounds are very rare, they are therefore not considered productive. This, I strongly disagree since V-N compounds exist in the language and it is a good source of new compounds. It is frequently used by the native speakers of Dagbanli in a natural conversation. Table XIV presents evidence of V-N compounds.

Table XIV. V-N Compounds in Dagbanli

Verb	Noun	Compound (V+N)	Gloss
bàɣɨ́sɨ́ ‘grind’	kʊ́ɣɨ́lɨ́ ‘stone’	bàɣɨ́sɨ́-kʊɣɨ́-lí	‘grinding stone’
dɨ́bʊ́ ‘eating’	ʃè-è ‘place’	dɨ́b-ʃè-è	‘eating place’
kàɾɨ́m ‘read’	dú-ú ‘room’	kàɾɨ́mbʊ́-dú-ú	‘reading room’
kàɾɨ́`m ‘read’	bì-á ‘child’	kàrɨ́m-bì-á	‘pupil’
vʊ́hɨ́m ‘rest’	dàbɨ́sɨ́lɨ́ ‘day’	vʊ́hɨ́m-dabɨ́sɨ́lɨ	‘rest day’

The data in Table XIV above show V-N compounds with a verb and a noun as its constituents. It shows how a verb occurs in initial position of a compound (e.g., baɣɨsɨ-kʊɣɨlɨ ‘griding stone’).

Headedness in Compounding

In compounding, the concept of headedness has played a central role forming the ‘basis for the distinction between endocentric compounds and exocentric compounds in its formation’ (Scalise & Vogel, 2010).

Studying Dagbanli data available for this paper, the paper observes that Dagbanli endocentric N-N compounds are right-headed and though there are left-headed and dual-headed N-N compounds. That is, the class of elements denoted by the compound will be a subset of the class of elements denoted by the head of the compound. For instance, bàɣɨ́sɨ́-kʊɣɨ́-lí ‘grinding stone, is a type of stone, the head. The formal head is the constituent that shares its formal features with the whole compound so that the whole compound has the same distributional properties as its formal head. This shows that bàɣɨ́sí ‘grind’ can occur where bàɣɨ́sí-kʊɣɨ́-lí is expected to occur without a drastic change in the meaning of the construction but the same cannot be said about the non-head constituent kʊɣɨ́-lí occurring alone where bàɣɨ́sí-kʊɣɨ́-lí is expected. This might suggest that endocentric compounds in Dagbanli are mostly likely to be right-headed compounds.

Some scholars (Booij, 2007; Dressler, 2006; Katamba, 1993) and others also distinguish between a semantic head, a syntactic head and a morphological head. Dressler (2006), for example, argues that the compound pickpocket has no semantic head because the referent is not named in the compound itself. Pick is the syntactic head because it selects pocket as its internal argument. The morphological head is pocket because when the compound is pluralized, the plural marker attaches to pocket (as in, pickpocket-s not pick-s-pocket). This can be argued in Dagbanli, for example, the compound nù-máhá ‘hand-wet’ similarly has no semantic head because the referent ‘thief’ is not named in the compound itself. It is observed that nù-ù ‘hand’ is the syntactic head but morphologically, máh-á ‘wets’ is the head because it takes the plural marker when the compound is pluralized, the plural marker is attached to the suffix noun as illustrated below:

Table XVa is grammatical because the plural marker –a is attached to the noun màhɨ́-lí while Table XVb is ungrammatical because the plural marker –hi is attached to the base noun nù-ù. Some exceptions exist in the language where speakers find it difficult to state what their plural forms are. The two common properties of these words would be that they are exocentric and they form complete clauses, describing the typical action or property of a particular object to the agent. One might therefore call them clausal nouns as this is illustrated below:

Table XV. Headedness in Dagbanli

a.	nù-ù + máh-á	→	nù-máhá
	hand.sg wet.pl		‘thief’
b.	nú-hí + màhɨ́-lí	→	*nú-hí màhɨ́lí
	hand.pl		‘wet.sg’

The example in Table XVI describes a flower, the marsh lily, which is said be a ‘beautiful woman (that) will not touch water’.

Table XVI. Clausal Noun in Dagbanli

páɣ-vɨ́Ɛl-lí	kù	ʃí-hí	kɔ́m
woman-nice	not	touch	water
‘Beautiful woman will not touch water’ (meaning water lily)

Following the two subclasses of bahuvrihi compounds types: the possessive and the non-possessive. The paper uses the possessive type and the non-possessive type to test exocentric bahuvrihi compounds in Dagbanli. Bahuvrihi is the type of compound in which the whole is not a hyponym of either constituent. Rather, it “expresses some salient facet of the denotatum” (Bauer, 2007, p. 56).

Observed from the data in Tables XVII and XVIII that both possessive bahuvrihi compounds and non-possessive bahuvrihi compounds are possible constructions in Dagbanli.

Table XVII. Possessive Bahuvrihi Compounds

Compounds	Constituents			Type
zʊ́ɣ-títá-lí ‘bighead’	zʊ́ɣ + títálí	Head + big	Person with a big head	N-A
zʊ́ɣ-bɨ́l-á ‘small head’	zʊ́ɣ + bɨ́l-á	Head + small	Person with a small head	N-A
tɨ́b-ցbándí ‘leather ears’	tɨ́b + ցbándí	Ear + leather	Person with big ears	N+N
zʊ́ɣ-pál-ó ‘bald head’	zʊ́ɣ + pál-ó	Head + plot	Bald head person	N+N

Table XVIII. Non-Possessive Bahuvrihi Compounds

Compounds	Constituents			Type
nànցbàn-jɨ́nɨ́ ‘mouth one’	nànցbà-nɨ́ + jí-ní	Mouth + one	‘unity’	N-N
nànցbàn-tɔ́ɣʊ́ ‘mouth bitter’	nànցbà-nɨ́ + tɔ́ɣʊ́	Mouth + bitter	‘hunger’	N-N
bɨ́nkɔ́b-zʊ́ɾɨ ‘animal tail’	bɨ́nkɔ́b-ցʊ́ + zʊ́-ɾi	Animal + tail	‘milk’	N-N
nànցbàn-málɨŋ ‘mouth sweet’	nànցbànɨ́ + málɨ́ŋ	Mouth + sweet	‘talkative’	N-A
nànցbàn-kpééní ‘mouth strong’	nàngbànɨ́+kpéé-ní	Mouth + strong	‘argument	N-A
jέl-kʊ́ɾ-lɨ́ ‘mouth old’	jέl +kʊ́ɾ-lɨ́	Problem + old	‘custom’	N-A
jέl-máŋ-lɨ́ ‘problem real’	jέl + máŋ-lɨ́	problem+ real	‘truth’	N-A

Compounds and Phrases

There are four ways of distinguishing compounds from phrases: “phonological, morphological, morpho-syntactic and semantic”. This paper discusses phonological looking at the assumption that the application of the four criteria is language-specific since every language has its own internal structure. Aikhenvald (2007, p. 25) observes, “unlike phrases, which consist of several phonological words, compounds often form one phonological word and thus have just one stress.” For the issue of compounds being stressed on the first of the two elements does not apply in Dagbanli because stress is not applicable in the language. Dagbanli is a tonal language and for that matter its compounds are marked by tone as seen in Table XIX.

Table XIX. Tone-Marked Compounds

Compounds	Element gloss	Meaning	Pattern
sá-kʊ́ɣá	rain + stones	‘hails’	N-N
jɨ́l-páɣá	house + woman	‘house wife’	N-N
kpáŋ-ցál-lɨ́	guineafowl + egg	‘guineafowl egg;	N-N
dá-jʊ́lɨ́	alcohol + pot	‘pot for drinking pito’	N-N
bɨ́n-pjέl-lɨ́	thing + white	‘white calico’	N-A
wáh-ʒjέɣʊ́	snake + red	‘type of snake’	N-A
tí-kpíl-lí	medicine + round	‘medical pill’	N-A
ցbán-pjέl-lɨ́	skin + white	‘western person’	N-A

The idea of using stress to distinguish between compounds and phrases does not apply in Dagbani as shown in the data because the language has no stress pattern as noted. For instance, the compound sá-páɣá which literally means ‘rain’s wife’ but actually means ‘dragonfly’, maintains the high tone of the first element making it different from the phrase sà–páɣá with a low tone on the first element.

Again, nominal noun compounds types may be used for different compounds as illustrated in Table XX for two nouns forming a phonological word and illustrated in Table XXI those for similar to possessive construction comprising two phonological words.

Table XX. Nominal Noun Compounds

Compound	Gloss	Nouns	Gloss	Pattern
bɨ́ŋ-káh-lí	‘unripe thing’	bɨ́ŋ + káh-lí	‘thing unripe’	N-N
dʊ́ŋ-kcᴐɣ-ʊ́	‘slim knee’	dù-ní + kcᴐʔ-ú	‘knee slim’	N-N
dʊ́n-bí-á	‘knee cap’	dù-ní + bí-á	‘cap knee’	N-N
láŋ-pjέl-lí	‘white net’	láŋ-á + pjέl-lí	‘net white’	N-N

Table XXI. Possessive Constructions in Dagbanli

Compound	Nouns	Meaning gloss	Meaning	Pattern
bɨ́n pjέl-lí	bini + pjέl-lí	Thing of white	‘calico’	N-A
kàɾɨ́m-zɔ́ŋ	kàɾɨ́m + zɔ́ŋ	Hall of reading	‘school’	V-N
nà-pá-á	nà-á + páɣ-á	Wife of chief	‘queen’	N-N
sà-kʊ́ɣɨ́-lí	sá-á + kuɣɨ́-lí	Rain of stones	‘hail’	N-N

I argue here that these compounds are formally indistinguishable from possessive constructions, such as nà-á-jílí ‘chief’s house.’ Note that the first example, bɨ́n-p^jέl-lí ‘calico’ in Table XXI is lexicalized. Olawsky (2002, p. 214) asserts that, “such noun–adjective constructions are particularly interesting since these adhoc combinations of nouns and adjectives in a nominal phrase containing an adjectival phrase are built in the very same way as lexicalised noun–adjective compounds.” This is a good reason to call these constructions ‘compounds’ in Dagbanli.

Conclusion

The paper examined some core issues in the study of compounding in Dagbanli focusing on typology and headedness. The paper described Dagbanli word as lexical items with CV syllable form without a coda (e.g., tú ‘insult’). The simplex form can be inflected for aspect (e.g., tí-já ‘giving’). The paper argued that compounds in Dagbanli consist of only two bases in which the final vowel of the first constituent is either reduced to a weak vowel or deleted due to its position within the compound. Based on the syntactic category of the constituents of a compound, the possible typology of compounds in Dagbanli are classified as Noun–Noun (N-N), Noun–Adjective (N-A), Noun–Verb (N-V) and Verb-Noun (V-N), which form nominals. The compounds consist of two nominal roots, the first element occurs without its class suffix, whereas the second part keeps its regular suffix, being the morphological head. On the part of headedness, the paper concluded that Dagbanli endocentric (N-N) compounds are dominated by right-headed (RH) though there are left-headed (N-N) compounds while all (N-A), (N-V) and (V-N) compounds are left-headed (LH).

Distinguishing compounds from phrases, the paper argued that compounds being stressed on the first of the two elements does not apply in Dagbanli and that there is no stress pattern in Dagbanli because Dagbanli is a tonal language. Hence, the idea of using stress to distinguish between compounds and phrases may not apply in Dagbanli. The paper argued that Dagbanli compounds supports the idea of compounds taking one inflectional marker on the final constituent but in a phrase, plural is marked on both components as it occurs in (N-N) noun phrase structure.