On the origin of some northern songhay mixed languages

2.1.3 Tense, Aspect, Mode

The TAM subsystems in all three NS languages are also of Songhay origin. The default TAM choice in most, especially spoken, discourse has no prefix. It functions as a completive aspect or, in stative verbs, something which is presently true. Table 6 demonstrates examples.

[INSERT TABLE 6 HERE]

Table 6: completive aspect, Northern Songhay languages^29
	Tagdal	Tadaksahak	Tasawaq
a.	irikoy. iri= koy 1p go ‘We left.’	arekoy. are= koy 1p go ‘We left.’	irikoy. iri= koy 1p go ‘We left.’
b.	iʒin-iri. i= ʒin =iri 3p grab 1p ‘They grabbed us.’	idʒin-are. i= dʒin =are 3p grab 1p ‘They grabbed us.’	idʒin-iri. i= dʒin =iri 3p grab 1p ‘They grabbed us.’
c.	adut hayni. a= dut hayni 3s pound millet ‘She pounded the millet.’	adut hayni. a= dut hayni 3s pound millet ‘She pounded the millet.’	adut hayni. a= dut hayni 3s pound millet ‘She pounded the millet.’
d.	ɣafur tonʒi. ɣa= fur tonʒi 1s throw stone ‘I threw a stone.’	aɣafur tondi. aɣa= fur tondi 1s throw stone ‘I threw a stone.’	ɣay fur tondi. ɣay fur tondi 1s throw stone ‘I threw a stone.’
e.	abarkaw anas. abarkaw a= nas calf 3s fat ‘The calf is fat.’	abarkaw anas. abarkaw a= nas calf 3s fat ‘The calf is fat.’	aʒemur anas. aʒemur a= nas calf 3s fat ‘The calf is fat.’

The prefix b-, the incompletive aspect, functions to indicate habitual actions; actions that were occurring in the past, which may or may not be the case in the present; and, in the case of non-stative verbs, actions that are occurring presently, as Table 7 shows.

[INSERT TABLE 7 HERE]

Table 7: incompletive aspect, Northern Songhay languages
Tagdal	Tadaksahak	Tasawaq
a. abhaŋga ǝnda-a. a= b- haŋga ǝnda =a 3s INC accompany with 3s ‘He hangs out with him.’	abhaŋga ǝnda-a. a= b- haŋga ǝnda =a 3s INC accompany with 3s ‘He hangs out with him.’	abhaŋga ǝnda-a. a= b- haŋga nda =a 3s INC accompany with 3s ‘He hangs out with him.’
b. ʒaɣʒi kulːu, abkoy yaːbu. ʒaɣʒi kulːu a= b- koy yaːbu day all 3s INC go market ‘He used to go the market every day.’	zaɣri kamil, abkoy suq. zaɣri kamil a= b- koy suq day all 3s INC go market ‘He used to go the market every day.’	ʒaɣʒi kulːu, abkoy yaːbu. ʒaɣʒi kulːu a= b- koy yaːbu day all 3s INC go market ‘He used to go the market every day.’
c. kala ɣabʒe takafar. kala ɣa= b- ʒe takafar before 1s INC speak French ‘I used to speak French before.’	kala aɣabdʒe takafart. kala aɣa= b- dʒe takafart before 1s INC speak French ‘I used to speak French before.’	kala ɣabsi takafart. kala ɣa= b- si takafart before 1s INC speak French ‘I used to speak French before.’
d. ɣabsǝrǝŋkat ɣan iman sa. ɣa= b- sǝrǝŋkat ɣa= n iman sa 1s INC brush hair 1s GEN soul DAT ‘I am brushing my hair.’	aɣabsǝrǝŋkat aɣan iman sa. aɣa= b- sǝrǝŋkat aɣa= n iman sa 1s INC brush hair1s GEN soul DAT ‘I am brushing hair.’	ɣabsiriŋkit ɣan iman sa. ɣa= b- iriŋkit ɣa= n iman sa 1s INC brush hair 1s GEN soul DAT ‘I am brushing my hair.’

{INSERT TABLE 8 HERE]

In Table 8 we see that the prefix m- or, ma- in Tasawaq, functions as the subjunctive or irrealis.

Table 8: irrealis in Northern Songhay languages
Tagdal	Tadaksahak	Tasawaq
a. irimkoy. iri= m- koy 1p IRR go ‘Let’s go. / We should go.’	aremkoy. are= m- koy 1p IRR go ‘Let’s go. / We should go.’	irimakoy. iri= ma- koy 1p IRR go ‘Let’s go. / We should go.’
b. irimwa. iri= m- wa 1p IRR eat ‘Let’s eat.’	aremŋwa. iri= m- ŋwa 1p IRR eat ‘Let’s eat.’	irimawa. iri= ma- wa 1p IRR eat ‘Let’s eat.’

[INSERT TABLE 9 HERE]

The prefix tǝ-, or in some cases bti- in Tasawaq, functions as the future tense (Table 9).³⁰

Table 9: future tense in Northern Songhay
Tagdal	Tadaksahak	Tasawaq
a. ɣatǝkoy. ɣa= tǝ- koy 1s FUT go ‘I will go.’	aɣatǝkoy. aɣa= tǝ- koy 1s FUT go ‘I will go.’	ɣatikoy. ɣa= ti- koy 1s FUT go ‘I will go.’
b. iritǝda-a. iri= tǝ- da =a 1p FUT do 3s ‘We’ll do it.’	aretǝda-a. are= tǝ- da =a 1p FUT do 3s ‘We’ll do it.’	iritida-a. iri= tǝ- da=a 1p FUT do 3s ‘We’ll do it.’
c. baŋgu atǝ ʃi ne. baŋgu a= tǝ- ʃi ne well 3s FUT not-be here ‘There will not be a well here.’	baŋgu atǝ ʃi ne. baNgu a= tǝ- ʃi ne well 3s FUT not-be here ‘There will not be a well here.’	baŋgu btisi ne. baŋgu bti- si ne Well FUT not-be here ‘There will not be a well here.’

Therefore, it seems clear that all three languages in question have similar inflectional Songhay morphologies. Table 10 below summarises the inflectional inventories of Northern Songhay and compares them with the inflectional inventories of Gao Songhay and Tayart.

[INSERT TABLE 10 HERE]

Table 10: Tense Aspect Mode in NS, mainstream Songhay and Berber
	Tagdal	Tadaksahak	Tasawaq	Tayart^31	Gao Songhay
Completive				-	
Incompletive	b-	b-	b-	-	ga/go^32
Future	tǝ-	tǝ-	tǝ-/bti-	-	ga-ti/go
Irrealis	m-	m-	ma-	-	ma-

Next we discuss their derivational subsystems of each of the Northern Songhay languages. As we will see, it is here that the differences between sedentary and nomadic will become most apparent.

2. 
   Issues in derivational morphology
   

The derivational sub-systems of each of the Northern Songhay languages are next for discussion. As we will see, it is here that the differences between sedentary and nomadic will become most apparent. Up to now, the structures of all three Northern Songhay languages surveyed have been fairly similar. However, when verb valence changes, the nomadic Northern Songhay languages – Tagdal and Tadaksahak – diverge from the sedentary Tasawaq by utilising Berber derivational morphology. In this section, we discuss the Berber causative in Tagdal and Tadaksahak (section 2.2.1), the Songhay causative in Tasawaq (section 2.2.2), then the reflexive (section 2.2.3) along with a syntactic construction with a similar function (section 2.2.4), and the passive voice prefixes (section 2.2.5), as well as a syntactic construction with similar function to the Passive (section 2.2.6). What we find is what Matras (2012) styled ‘bilingual diffusion’.

1. 
   Causative (nomadic Tagdal, Tadaksahak)
   

In the nomadic varieties, when the verb root is of Berber origin, as in the examples below,³³ the default strategy when forming causatives is to add the prefix s- , or its allophones ʃ-, z- or ʒ- (see Table 11).

[INSERT TABLE 11 HERE]

Table 11: causative verbs in nomadic Northern Songhay languages^34
Verb root, Tadaksahak	Verb root, Tagdal	Causative verb, Tagdal/Tadaksahak
a. yirzǝg ’move about’	ǝrzǝg ‘move about’	zǝrzǝg ‘cause to move about’
b. yirkǝb ‘pull’	ǝrkǝb ‘pull’	sǝrkǝb ‘cause to pull’
c. yilkǝd ‘overtake’	ǝlkǝd ‘overtake’	sǝlkǝd ‘cause to overtake’
d. yiklu ‘take a midday rest’	ǝklu ‘take a midday rest’	sǝklu ‘cause to take rest’
e. yiɣli ‘surround’	ǝɣli ‘surround’	sǝɣli ‘cause to surround’
f. yiħkǝm ‘rule, judge’	ǝħkǝm ‘rule, judge’	sǝħkǝm ‘cause to judge’

Below in Table 12, we include data from Təwəlləmət and from Zarma. Our purpose here is to demonstrate that the causative prefix s- and the vocabulary above are of Berber origin, as well as to demonstrate how mainstream Songhay languages typically form causatives.

[INSERT TABLE 12 HERE]

Table 12: causative verbs in mainstream Songhay, Berber
Tǝwǝllǝmmǝt	Zarma^35
a. ǝrzǝg ‘move about’ zǝrzǝg ‘cause to move about’	yota ‘move about’ yotandi ‘cause to move about’
b. ǝrkǝb ‘pull’ sǝrkǝb ‘cause to pull’	tʃan ‘pull’ tʃandi ‘cause to pull’
c. ǝlkǝd ‘overtake’ sǝlkǝd ‘cause to overtake’	bisa ‘overtake’ bisandi ‘cause to overtake’
d. ǝklu ‘take a midday rest’ sǝklu ‘cause to take a rest’	foya ‘take a midday rest’ foyandi ‘cause to take a midday rest’
e. ǝɣli ‘surround’ sǝɣli ‘cause to surround’	wi ‘surround’ windi ‘cause to surround’
f. ǝxkǝm ‘rule, judge’ sǝxkǝm ‘cause to rule, judge’	meera ‘rule, judge’ meerandi ‘cause to rule, judge’

Therefore, it seems fairly clear that the causative s- is of Berber origin and only affixes onto verb roots of Berber origin in Tagdal and Tadaksahak. But what if the verb root is of Songhay origin? In nomadic Tagdal and Tadaksahak, if the verb root is of Songhay origin and a causative is required, an already-causativised Berber verb suppletes the Songhay verb root,³⁶ as in the examples below in Table 13 below.

[INSERT TABLE 13 HERE]

Table 13: suppletion of Songhay verb roots, nomadic Northern Songhay languages
Verb root of Songhay origin		Causative form
a. koy ‘go’		sǝglu ‘cause to go’
b. goːra ‘sit’		sǝɣima (Tad.)/sǝqima (Tag.) ‘make sit’
c. qoq ‘dry’		sǝɣǝr ‘cause to dry’
d. nin ‘drink’		ʃǝʃǝw ‘give someone a drink’
e. zumbu ‘descend’		zǝzǝbǝt ‘cause someone to come down’

2. 
   Causative (sedentary Tasawaq)
   

In Tasawaq, the sedentary Northern Songhay language, the default causative consists of adding the suffix –nda to verb roots of Songhay origin.³⁷ This is the only recorded productive derivational verbal affix in Tasawaq. The following forms in Table 14 are some attested causative verbs in Tasawaq.
[INSERT TABLE 14 HERE]

Table 14: Songhay causative verbs, Tasawaq, sedentary Northern Songhay language
Verb root	Causative verb
a. taŋɣari ‘lie’	taŋɣarinda ‘cause to lie’
b. sigːiaːre ‘lie down’	sigːiaːrenda ‘cause to lie down’
c. wa ‘eat’	wanda ‘cause to eat’
d. zumbu ‘come down’	zumbunda ‘cause to come down’
e. nin ‘drink’	ninǝnda ‘cause to drink’

Some Berber causatives do exist in Tasawaq. The following forms, in Table 15, are a few that we have run into in natural speech, during the course of conversations with Tasawaq speakers.

INSERT TABLE 15 HERE]

Table 15: some examples of Berber causatives in Tasawaq
a. sǝṭkǝl	‘cause to lift up, take’
b. sǝrkǝb	‘cause to pull’
c. zǝzzǝbǝt	‘cause to come down’

The existence of these Berber causative forms in sedentary Tasawaq creates difficulties. ³⁸ First, their non-causative equivalents are not verbs in Tasawaq. For example, the verb ǝṭkǝl* ‘lift up / take’ is indeed a verb in nomadic Tagdal and Tadaksahak. However, in Tasawaq the verb is dʒin, with the causative dʒinǝnda ‘cause to take’ being very possible. Likewise, ǝzzǝbǝt* ‘come down’ is not a verb in either Tasawaq, nor in the other NS languages; the verb ‘come down’ is actually zumbu, with the causative zumbunda ‘cause to come down’ in Tasawaq (Alidou 1988: 50). Further, despite the existence of Berber causatives in Tasawaq, neither the Passive nor the Reflexive prefixes are possible even with the same Berber roots.³⁹ In other words, it seems that the Berber causative forms might be cases of borrowed vocabulary, and not productive like in the nomadic languages.

Another, more serious, difficulty in Tasawaq is the presence of a few y- verbs (Kossmann 2007b), underived Berber cognates of those possible in Tagdal and Tadaksahak (see Table 11 above). If these behave in Tasawaq as they do in the nomadic NS languages, presumably they would take the Berber causative prefix s-. Admittedly, however, this is not a question that Kossmann was addressing in Kossmann 2007b. So we do not know how these would derive.

The problems with the underived y- verbs in Tasawaq that we ourselves observed are:

1) They were always produced by multilingual Tasawaq speakers;

2) We have only observed instances of the Berber causative, never the reflexive or passive prefixes (see sections 2.2.3 and 2.2.4 below). So it is not the full Berber derivational sub-system, as occurs in the nomadic languages;

3) Alidou (1988) did not mention y- verbs in her thesis; in fact, she states that the Songhay causative suffix -nda “…combines with any verb.” (1988: 50)⁴⁰ So either y- verbs were not enough of a factor in the grammar to include when she was writing her thesis in 1988, or perhaps they are a newer phenomenon in the language.

4) Finally, we observed that often y- verbs occurred in natural speech (again, always produced by multilingual speakers), even in cases where a perfectly good, more common, verb of Songhay origin existed in Tasawaq.⁴¹

Regardless of the very notable exceptions above, in the case of the causative, the nomadic and sedentary branches of Northern Songhay have very different default strategies. In Tagdal and Tadaksahak, if the verb root is of Berber origin, the strategy is to add the prefix s- or one of its allomorphs. If the verb root is of Songhay origin, an already-causativised Berber verb root suppletes the Songhay root. The only exception we have found is the case of the Songhay causative kanda ‘cause to fall’ in Tagdal. On the other hand, in Tasawaq the default causative morpheme, the above exceptions aside, is the suffix –nda affixed onto a verb root of Songhay origin. The few cases of Berber causatives in Tasawaq seem to be non-productive, fixed, forms.

3.