Two Types of Temporal When Clauses in Hausa [1]
Mahamane L. Abdoulaye*
Hausa has simple temporal clauses, introduced by the conjunction dą when, that many researchers derive from temporal relative clauses introduced by lookącin dą time that, through the deletion of the pseudo head word lookącii time. This paper shows that this analysis, however natural it may appear, is inadequate. Indeed, the two temporal clauses have different uses, which would not be surprising if simple temporal clauses are derivationally independent from temporal relative clauses, as proposed in this paper.
There are in Hausa
two types of temporal when clauses.
The first type has the structure of relative clauses while the second type
shows no immediate connection to relative clauses. The two types of temporal
clauses are illustrated in the following:
(1) a. Sun kaawoo ma-ną zģyaarąa lookąci-n dą
3P.CPL bring to-1P visit time-DF that
mu-kče kąaląacii.
1P-RI meal
They visited us when we
were having a meal.
b. Sun kaawoo ma-ną zģyaarąa dą
3P.CPL bring to-1P visit when
mu-nąa kąaląacii.
1P-IPV meal
They visited us when we
were having a meal.
The temporal clause
in (1a) is structurally a relative construction where the relative clause introduced
by dą seems to modify the apparent
head word lookącii time, moment,
period. Sometimes, instead of lookącii,
an alternative word may appear such as saąa,
sąaģdii, yąayii, etc., all meaning time, moment, etc. In (1b) by contrast,
the temporal clause is introduced by the particle dą alone which is here translated as when. Furthermore, the two
temporal clauses differ in the fact that the temporal relative clause can
optionally take the relative marking, i.e., alternate forms of the perfective
and imperfective that appear in relative clauses and out-of-focus clauses of
focus and wh question constructions
(see among others Bearth 1993; Hyman and Watters 1984; Schachter 1973). For
this reason, the temporal relative clause in (1a) has the relative imperfective
mu‑kče 1P‑RI contrasting with the regular imperfective mu‑nąa
1P‑IPV found in the simple temporal clause in (1b).
Nearly all
descriptions of Hausa temporal clauses claim or assume that temporal relative
clauses, especially the ones headed by the word lookącii time, are the
source of simple temporal clauses introduced by dą. The derivation would
involve the deletion of the word lookącii time or its equivalents (cf.
Bagari 1976/87: 117; Jaggar 2001: 624; Newman 2000: 556; Tuller
1986: 113). In fact, for most authors (cf. Jaggar 2001: 624, 629),
the lookącii temporal relative clause derives a whole series of temporal
clauses introduced by phrasal subordinators involving the particle dą,
such as: (lookącin) dą (time) when, sai (lookącin) dą till
(time) when, tun (lookącin) dą since (time) when, etc. The claim that
temporal relative clauses are the source of simple temporal clauses is usually
based on examples where the word lookącii time seems optional, as
illustrated next (cf. also Bagari 1976/87: 117; Watters 2000: 223):
(2) a. Naa san Abdł (lookąci-n) dą ya-nąa yaarņo.
1s.CPL know Abdu time-DF DA 3MS-be child
I know Abdu (at the time)
when he was a child.
b.
Yāaraa sun ga sarkii (lookąci-n) dą su-ką
children 3p.CPL see emir time-DF DA 3P-RP
shģga gąrii.
enter town
The children saw the emir
when they visited the town.
The children saw the emir
when they were entering the town.
In the sentences in
(2), the presence or absence of the word lookącii
time has no consequence on the meaning of the sentences. In (2b) for example,
with or without lookącii, there is an
ambiguity between the interpretations the children saw the emir at the exact
moment when they entered the town and the children saw the emir when they
were visiting the town. It thus seems completely natural to derive the simple
temporal clause from the more complex temporal relative clauses through the
deletion of the word lookącii. In
fact, this process is thought to be general and, according to Wald (1987:
509n5), many West African languages commonly use a relative conjunction (such
as dą in Hausa) as a conjunction
introducing temporal when clauses. [2]
The aim of this
paper is to show that the assumption of a systematic derivation of simple
temporal clauses from temporal relative clauses cannot be maintained when one
closely examines the uses of the two types of clauses. The paper thus presents
a series of indications suggesting that the simple temporal clauses are not
derived from temporal relative clauses.
As will be seen in
due course, in trying to establish the independence of simple temporal clauses,
this paper retraces the development of both types of clauses using the
grammaticalization framework. Normally, a grammaticalization process, in a specific
context, turns a lexical or derivational item into a grammatical marker, or a
grammatical marker becomes more grammatical (cf. for example Hopper and
Traugott 1994: 2). In this process, the original lexical item becomes
progressively eroded, both at the phonological and semantic levels. However,
there is another process that also falls under the domain of
grammaticalization, where an entire construction undergoes expansion in new
contexts and/or acquires new functions while becoming syntactically more
integrated and less flexible. This type of grammaticalization has been
discussed in, among others, Givón (1990: 651); Güldemann (2003: 183);
Heine and Reh (1983: 34); Himmelmann (1997); and Hopper and Traugott
(1994: 167ff). In this paper, we will see both types of changes. Indeed,
simple temporal clauses stem from the grammaticalization of the dą, which evolved from an existential
predicate, through a comitative and instrumental marker, to a temporal
conjunction. Temporal relative clauses on the other hand are derived from
typical head modifying relative clauses and have a frozen structure
characteristic of grammaticalized constructions.
The paper is
structured as follows. Section 2 presents the relationship between temporal
relative clauses and regular relative clauses. Section 3 sketches a possible
development scenario for simple temporal clauses that makes no appeal to
relative constructions. Finally, Section 4 discusses other differences
between the two types of temporal clauses, in particular their semantics and
their interaction with times/aspects/modes (TAM), differences that show the
derivational independence of simple temporal clauses vis-ą-vis temporal
relative clauses.
It has naturally
long been evident that adverbial clauses in general may in some languages be
closely connected to relative constructions (cf. Lehmann 1984: 320;
Thompson and Longacre 1985: 178). This section shows that temporal
relative clauses in Hausa can be derived from relative clauses where the head
word lookącii time has a true
referential reading.
Like any noun phrase
functioning in a sentence (as subject, direct object, etc.), the word lookącii can be modified by at least
four kinds of relative clauses. One of the two fundamental contrasts in Hausa
relative clauses distinguishes relative clauses with a complex structure
involving a copular predicate kč(e)
be and simpler, reduced relative clauses that have no copula. These two types
of relative clauses are illustrated in the following (cf. also Newman
2000: 540):
(3) a. Ką ąuni lookąci-n dą ya-kč yaa daacče.
2MS.SUB aim time-DF that 3MS-be 3MS.CPL fit
You must/may aim at the
really appropriate moment.
b. Ką ąuni lookąci-n dą ya daacče.
2MS.SUB aim time-DF that 3MS.RP fit
You must/may aim at the
appropriate moment.
In the sentences in
(3), the main clause is ką ąuni
[lookącii
] you should target [a time
] where the word lookącii functions as direct object in
the main clause. In (3a), the relative clause is introduced by the conjunction dą that and is made up of an
impersonal copular predicate ya-kč
it be and the relative clause proper. We will assume that the relative clause
proper here functions as a complement to the copula kč. By contrast, in (3b), the relative clause is reduced in
structure and is introduced only by the conjunction dą that. As may be seen, the reduced relative clause in (3a) has
the relative perfective marker ya
3MS.RP replacing the regular completive marker yaa 3MS.CPL found in the copular relative clause in (3a). It may
be noted that reduced relative clauses as illustrated in (3b) are more
frequent, while the copular relative clauses illustrated in (3a) have an
associated emphatic reading, as indicated in the translation.
The second fundamental
contrast in Hausa relative clauses is the presence or absence of a relative
pronoun introducing the relative clause. Indeed, the relative clauses
illustrated in (3) all have variants with a relative pronoun, as illustrated in
the following:
(4) a. Ką ąuni lookącii wa-n-dą ya-kč
2MS.SUB aim time one-DF-that 3MS-be
yaa daacče.
3MS.CPL fit
You must aim at the really
appropriate moment.
b. Ką ąuni lookącii wa-n-dą ya daacče.
2MS.SUB aim time one-DF-that 3MS.RP fit
You must aim at the appropriate
moment.
c. wa-n-dą ya-kč yaa daacče
one-DF-that 3MS-be 3MS.CPL fit
the one that is really
appropriate.
Relative clauses
without a relative pronoun, as illustrated in (3), are usually taken to be more
basic (cf. Newman 2000: 540). Indeed, relative clauses introduced by a
relative pronoun, as illustrated in (4a‑b) are historically secondary and
stem probably from the grammaticalization of free relative clauses headed by
the indefinite pronoun wąa one, as
illustrated in (4c). The free relative clauses are now alternating with the
original dą relative clauses in
modifying a head noun. In (4a‑b), the sequence made up of the indefinite
pronoun and the conjunction (i.e., wa-n-dą
the one-masc. that) is reinterpreted as a relative pronoun wandą who-masc.. The relative pronoun
is written as one word in Hausa orthography and the other forms are: waddą/waccč who-fem., wadąndą
who-plur. (theses pronouns in fact have a variable tone pattern, see Jaggar
2001: 528; on the development of relative pronouns in general, see among
others Givon 1990: 657; Lehmann 1984: 389; van der Auwera and
Kučanda 1985: 927, 953).
Data (3‑4)
illustrate the noun lookącii time
functioning as direct object of the main clause and this noun can naturally
also assume other syntactic functions in the main clause. In particular, lookącii can also function as a temporal
adverb in the main clause. However, in such case, lookącii and its accompanying relative clause have a double
interpretation, as seen in the next examples:
(5) a. Bąakii sun zoo lookąci-n dą
visitors 3P.CPL come time-DF that
mu-kče cīn ąbinci.
1P-RI eat food
The visitors came at our
meal time.
The visitors came while we
were eating.
b. Bąakii sun zoo jiyą/ karfče biyu.
visitors 3P.CPL come yesterday/ oclock two
The visitors came
yesterday/ at 2 oclock.
c. Lookącī-n nan nče bąakī-n su-ką zoo.
tile-DF that cop. Visitors-DF 3P.RP come
It was at that moment that
the visitors came.
In the first meaning
of (5a), lookącii is referential and
designates a precise moment. In this interpretation, the people visited do not
need to actually be eating. In (5a) then lookącii
fulfills a function in the main clause and is comparable to simple temporal
adverbs such as jiyą yesterday,
etc., as illustrated in (5b). The difference between the two types of adverbs
is that the word lookącii must
necessarily be specified by a relative clause to have a reference, unless this
reference is clear from context. For example, in (5c) the reference of lookącii is would be specified in the
preceding discourse. In the second meaning of (5a), the word lookącii time is not referential and
its interpretation seems to be closely tied to the action described in the
apparent relative clause (the action of eating). In particular, the action of
eating refers to a specific event (unlike in the first interpretation), which
so determines the temporal reference. In this case, it is the entire apparent
relative construction that acts as a temporal adverbial clause in the main clause.
That is, since lookącii is not
referential (or is lexically empty, cf. Lehmann 1984: 320), one no longer has
the construction Head lookącii +
relative clause. Indeed, many relative constructions with the word lookącii can only be interpreted as adverbial
clauses, in particular when they refer to single occurrence events, as
illustrated in the following:
(6) a. Lookąci-n dą su-nąa fadąa nee łbā-n
time-DF that 3P-IPV fight cop. father-DF
naa-sł ya fitoo.
of-3P 3MS.RP come.out
It is when they were
fighting that their father came out.
b. Lookąci-n dą Saanii ya buudoo koofąa
time-DF that Sani 3MS.RP open door
sai su-ką ruugąa.
then 3P-RP run
It is when Sani opened the
door that they fled.
In sentences (6),
the action described in the temporal relative clause is a single occurrence
event, whether the tense/aspect is imperfective, as in (6a), or perfective, as
in (6b). In such cases, the event in the temporal relative clause can be
interpreted as causal, leading to the event described in the main. Indeed, in
(6a‑b), the relation between subordinate and main clause event can be
temporal (fortuitous) or causal (i.e., respectively, the father came out in
order to see what was going on and the children ran away from Sani; cf. also
the discussion of data (16) below).
The
functional/semantic changes observed in data (5‑6) can be naturally
apprehended in the grammaticalization framework whereby the regular relative
clauses would be the source of the specialized temporal relative clauses.
Indeed, there are indications that the semantic changes are accompanied by
formal changes as well. As expected in a grammaticalization process, the
derived temporal relative clauses are indeed more reduced and less flexible
syntactically. We have at the beginning of this section seen that one contrast
opposes copular and reduced relative clauses (cf. discussion of data (3)) while
another contrast opposes relative clauses that have a relative pronoun and
those that have no relative pronoun (cf. discussion of data (4)). It happens
that temporal relative clauses have no variants with the copula ‑kč and cannot take a relative
pronoun, as illustrated in the following data:
(7) a. *Lookąci-n dą ya-kč su-nąa fadąa...
time-DF that 3MS-be 3P-IPV fight
When they were fighting
b. *Lookącii wa-n-dą su-kče fadąa...
time one-DF-that 3P-RI fight
When they were fighting
c. *Lookącii wa-n-dą ya-kč su-nąa fadąa...
time one-DF-that 3MS-be 3P-IPV fight
When they were fighting
Examples (7a‑c)
show that temporal relative clauses cannot, respectively, have the copula kč, the relative pronoun, or both
features at the same time. It may noted that copular relative clauses have an
emphatic connotation, which would apparently apply to lookącii time only if it is referential. At the beginning of the
section it was suggested that relative clauses with a relative pronoun are a
secondary development and incorporate a free relative clause. It appears then
that only reduced relative clauses without a relative pronoun specialized to
become temporal relative clauses, the more complex relative clauses (relative
clauses with copula kč and/or a
relative pronoun) are incompatible with the adverbial function.
In relative clauses
and temporal relative clauses, the particle dą
is generally taken to be a purely grammatical marker, i.e., a subordination
conjunction without semantic load. By contrast, the same particle in simple
temporal clauses is translated by all writers as the conjunction when (cf.
for example Jaggar 2001: 606, 624; Newman 2000: 556; cf. also data
(1b) above). Nonetheless, such translation is only an approximation and is
context bound since dą appears in
other temporal expressions where it is not exactly translatable as when. This
is illustrated in the following (cf. also Jaggar 2001: 650; Wolff
1993: 440):
(8) Abdł yaa zoo dą saafe/
Abdu 3MS.CPL come during early.morning/
(dą) karfče takwąs.
at oclock eight
Abdu arrived early in the
morning/ at 8 oclock.
As seen in (8), the
particle dą can, obligatorily or
optionally, accompany certain temporal adverbs such as the times of day, prayer
moments, hours, seasons, etc. (however, some temporal adverbs such as maakņo/saatii week, the months and the
years such as Maarģs march, 1999, bana
this year, bąara last year, and bądi
next year, etc.- do not take dą).
As indicated, in contexts such as (8), dą
can be translated as at, in/during, which shows that the temporal particle is
in fact semantically complex and the meaning of when may be a derived
meaning.
In this regard,
there are indications showing that particle dą
when may have derived from the comitative/instrumental preposition dą with. In fact in certain temporal
uses, the comitative semantics of dą
is quite explicit, as illustrated next:
(9) a. Abdł yaa zoo dą wuri.
Abdu 3MS.CPL come with margin
Abdu came very early (i.e.,
with space, margin).
b. Ciiwņ-n nān yaa zoo dą dąamanaa.
sickness-DF this 3MS.CPL come DA rainy.season
This disease came with the
rainy season.
This disease came during
the rainy season.
In (9a), the adverb wuri comes very likely from wurii place, space and the expression dą wuri literally means with (time)
room/margin. Data (9b) however may more clearly show the semantic/functional
shift from comitative dą to temporal dą. Indeed, dą in (9b) can mean with, as in the first interpretation, or
during, as in the second interpretation. One may assume that once particle dą took up the function of introducing
temporal adverbs, the new function spread to events expressed in finite or non
finite clauses. This is illustrated in the following (example (10a) adapted
from Hiskett 1971: 78 and (10b) from Moussa-Aghali 2000: 8; cf. also
Jaggar 2001: 635 for similar examples):
(10) a. Dą gaanąawaa dą kau rąsuwaa taa-są.
on meeting on indeed dying of-3MS
He died as soon as they
greeted.
b. Dą jī-n haką sai uwaa-taa
on hearing-of this then mother-of.1S
ta buushče dą dąariyaa.
3FS.RP blow with laughter
On hearing this, my mother
laughed.
c. Dą ta ji haką sai uwaa-taa
when 3FS.RP hear this then mother-of.1S
ta buushče dą dąariyaa.
3FS.RP blow with laughter
When she heard this, my
mother laughed.
In (10a), particle dą introduces two verbal nouns and is probably
the same preposition found introducing nominals in data (8‑9). The
construction dą + verbal noun +
(complement) is quite frequent in Hausa and can appear even in frozen
expressions (such as dą faarąawaa dą iyąawaa early talent, lit. on starting (is) on knowing how; cf. further
examples in Newman 2000: 44). For this reason, the construction is
frequently an alternative to finite temporal clauses, as seen in (10b‑c)
where the two sentences are essentially equivalent (although (10b) is also used
to convey the as soon as she heard this
reading). The shift from a
preposition to a subordinating conjunction seen in (10b‑c) characterizes
the development of many particles in Hausa (cf. sai Abdł only Abdu (can
do something) and sai kaa jee can only [if] you go there (can you
achieve something)). Other particles introducing a noun phrase or a finite
clause are: baayan after (from baayaa
back), koo even, kąafin before, tun since, etc.
(cf. Schachter 1985: 51; Wolff 1993: 449). It is clear that examples
(8‑10) evidence some temporal uses of dą that are unrelated to lookącin
dą relative clauses. The proposal that temporal conjunction dą stemmed from comitative/instrumental dą is hence a viable alternative to the lookącii deletion analysis.
In the previous two
sections, we saw that simple temporal clauses and temporal relative clauses may
have different origins. This section shows that the two types of clauses also
differ with regard to their use in ways that cannot be accommodated in the
framework of the lookącii deletion
analysis. Indeed, the two types of temporal clauses do not express the same
range of temporal when relations and
behave differently with regard to the tense/aspect paradigms.
In the general
linguistic literature (cf. Michaelis 2006, Vlach 1981) there are essentially
two types of temporal relations between the event in a temporal when clause and the event in its main
clause. Regarding English for example, Vlach (1981) considers that the relation
is overlapping when one of the event refers to a state (i.e., when the event is
durative). By contrast, the relation is consecutive when both events in the
subordinate and main clause are non durative. The overlapping relation is well
illustrated for both types of Hausa temporal clauses, as seen in data (1‑2)
where one of the events is durative. The consecutive relation with two punctual
events is also well illustrated in (6c) for the temporal relative clauses and
in (10c) for the simple temporal clauses. However, in Hausa the interpretation
of the temporal relation may not in fact depend on the durativity of the events
per se. Indeed, the two types of temporal clauses may in some cases have an
overlapping or consecutive reading when both subordinate and main clauses
describe punctual events. This is illustrated in the following:
(11) a. (Lookąci-n) dą ya taashģ, sai
time-DF DA 3MS.RP stand then
ya habrč kwaanņ-n ruwā-n.
3MS.RP kick bowl-of water-DF
When he was standing up, he
kicked over the water bowl.
When [after] he stood up,
he kicked over the water bowl.
b. (Lookąci-n) dą ya diroo, sai
time-DF DA 3MS.RP jump then
ya karč kafąa/ sąndā-r.
3MS.RP break leg/ stick-DF
When he jumped, he broke
his leg/the stick.
In (11), the
presence of the word lookącii has no
effect on the interpretation of the sentences so that all interpretations apply
both to simple temporal clauses and temporal relative clauses. In (11a), the
sentence is ambiguous between an overlapping and a consecutive interpretation.
This seems to hinge on the durative vs. punctual perspective taken for the
event in the temporal clause. Sentence (11b), too, is ambiguous between an
overlapping and a consecutive reading but in way different from (11a). While in
the first interpretation of (11a) the kicking of the bowl may happen anytime
during the process of standing up, in (11b) the breaking of the leg can only
happen, normally, at the end of the jumping process. It is clear that there is
here a minimal overlapping of the events. Depending on the nature of the
object, sentence (11b) can also have a consecutive reading (jumping and then
breaking a stick). The various types of temporal when relations seen so far between subordinate and main clause
event can be diagrammed as in the following:
(12) Full/partial overlapping -- minimal
overlapping -- consecution
As illustrated in
the examples, both types of temporal clauses can express all three relations in
(12). However, there is one further type of consecutive when relation that can be expressed only by simple temporal
clauses. In this temporal relation, the close sequencing of the events is
stressed or particularly explicit. The stressing of the close sequence relation
is illustrated next:
(13) a. (Lookąci-n) dą su-ką zoo, sai mu-ką
time-DF DA 3P-RP come then 1P-RP
ci ąbinci.
eat meal
When [once] they arrived,
we then ate.
b. Koo (*lookąci-n) dą su-ką zoo, sai mu-ką
even time-DF DA 3P-RP come then 1P-RP
ci ąbinci.
eat meal
As soon as they arrived, we
ate.
In (13a), the
closeness of the consecutive relation is not stressed and both types of
temporal clauses are possible, as shown by the optionality of lookącii time. In (13b), the sentence
is introduced by koo even, which is
an emphatic particle (cf. König 1991), stressing the close sequencing of the
events. One notes that in this case, the word lookącii is not possible. Another case of explicit close sequencing
is illustrated next:
(14) Ta tąmbąyi Saanii. (*Lookąci-n) dą ya
3FS.RP ask Sani time-DF DA 3MS.RP
baa tą, sai ta tąfi.
give 3FS then 3FS.RP go
She asked Sani [for sth.]. When he gave [it to] her, she went.
In (14), which could
be a piece of narrative, the close consecutive relation is explicit, i.e., in
the context of asking, the woman left as soon as she was given something. Here
too, the word lookącii is not
possible. The narrative context in fact provides a further illustration of a
close consecutive relation that excludes the temporal relative clauses. This is
seen in the following:
(15) Su-ką fģta. (*Lookąci-n) dą su-ką fģta,
3P-RP go.out time-DF DA 3P-RP go.out
sai ta rufč koofąa.
then 3FS.RP close door
They then left. Once they
exited, she then closed the door.
Narratives, by
definition, relate events that are sequenced and single occurrence (Adam
1994: 92‑105), typically in a close succession. In Hausa, a frequent
narrative technique is to repeat a previous event in a temporal clause, before
chaining up with the next event. In (15), the event of getting out is
presented. Then the same event is repeated in the subordinate clause and linked
with the next new event. This repetition in fact stresses the close
connectedness of the events, as shown in the translation. One notes that here,
too, the word lookącii is not
possible. In fact, the repeated event can explicitly be introduced by baayan after or a hybrid conjunction baayan dą after that/when, instead of dą alone. This is illustrated in the
following:
(16) a. Su-ką fģta. Baayan (*Lookąci-n) dą su-ką fģta,
3P-RP go.out after time-DF DA 3P-RP go.out
sai ta rufč koofąa.
then 3FS.RP close door
They then left. After they
exited, she then closed the door.
b. Su-ką fģta. Baaya-n sun fģta, sai
3P-RP go.out after 3P.CPL go.out then
ta rufč koofąa.
3FS.RP close door
They then left. After they
exited, she then closed the door.
In (16a), the first
event of the passage is repeated and introduced by baayan dą after that/when and lookącii
cannot be used. Baayan is a
preposition and conjunction meaning behind, after (cf. baayan iccče behind the tree, baayan
salląh after the festival). As seen in (16b), baayan can appear alone and mark the close sequence relation. [3]
Because of this
ability of simple temporal clauses to mark close consecutive events, they
usually have, given the appropriate context, more causative implication than
temporal relative clauses. This is illustrated in the following:
(17) a. Lookąci-n dą su-nąa fadąa nee łbā-n
time-DF that 3p-IPV fight cop. father-DF
naa-sł ya fitoo.
of.3P 3MS.RP come.out
It is when they were fighting that their father
came out.
b. Dą su-nąa fadąa nee łbā-n naa-sł
when 3p-IPV fight cop. father-DF of.3P
ya fitoo.
3MS.RP come.out
It is when they were
fighting that their father came out.
In the discussion of
data (6a) above, we said that temporal relative clauses can have causative
implications. Nonetheless, when asked to contrast the sentences in (17), most
Hausa speakers interpret (17a) [= (6a)] as implying that the father came out
inadvertently to find the fighting going on while (17b) is taken to imply that
the father came out on purpose (say upon hearing that a fight is going on). It
should be noted that both clauses primarily have a temporal function and, with
some main verbs, may or may not have a causative implication. For example if
the verb fitoo come out is replaced
by zoo come, then both (17a‑b)
would have a simple contingency reading (that is, if the father came from work
or from somewhere not knowing about the fight). Conversely, if, instead of a
father coming out, the main clause describes the police making arrest, then
both temporal clauses would in this context naturally get the causative
implication (that is, the fighters were arrested because of their fighting).
To summarize, simple
temporal clauses and temporal relative clauses do not have the same uses. In
the lookącii deletion analysis, which
derives simple temporal clauses from temporal relative clauses, this difference
in use will have to be explained in one way or another. By contrast, if the two
clauses developed independently (cf. Sections 2 and 3), then one can
expect them to have different properties. Next we see some further differences
between the two clauses.
This subsection
explores the tense/aspect paradigms that can appear in simple temporal clauses
and in temporal relative clauses, as well as the temporal interpretations of
the paradigms (i.e., whether they can refer to past or future events).
Table 1 presents the tense/aspect possibilities with temporal lookącin
dą relative clauses.
Table 1: TAM paradigms and their
interpretations in temporal relative clauses (with 3rd person plural
su‑ and verb fģta go out)
|
|
Past |
Future |
|
Imperfective: lookącin dą sunąa fģtaa |
Yes |
Yes |
|
Relative Imperfective: lookącin dą sukče fģtaa |
Yes |
Yes |
|
Relative Perfective: lookącin dą suką fģta |
Yes |
Yes |
|
Future I: lookącin dą zaa
sł fģta |
Yes |
Yes |
|
Eventual: lookącin dą sukąa fģta |
Yes |
Yes |
Hausa has about ten
TAM paradigms (cf. Abdoulaye 2008) but Table 1 shows that only five of
them can appear in temporal relative clauses: the regular imperfective, the
relative imperfective, the relative perfective, the future I, and the
eventual. Temporal relative clauses cannot take the habitual, the
future II, the subjunctive, etc. It should be noted that a relative clause
containing the habitual aspect can modify the word lookącii, time, but in this case the word lookącii receives a referential
interpretation only (i.e., the clause would not have an adverbial use; cf. lookącin dą sukąn fģta the time/moment
when they usually go out). Also, only clauses describing one-time events are
considered in Table 1 since they allow the adverbial use, while clauses
with recurrent events would tend to have a referential lookącii head. It may also be noted that a temporal relative clause
takes the regular or the relative imperfective with the same meaning, as
indicated in the introductory section.
The second
significant aspect of Table 1 is that all admissible tense/aspect
paradigms can refer to the past or the future, given an appropriate context (as
determined by the main clause). This is illustrated in the following for the
relative perfective:
(18) a. Lookąci-n dą bąakii su-ką zoo, an
time-DF that visitors 3P-RP come imp.CPL
baa sł tąabarmaa.
give 3P mat
When the visitors came,
they were given a mat.
b. Lookąci-n dą bąakii su-ką zoo, ą
time-DF that visitors 3P-RP come imp.SUB
baa sł tąabarmaa.
give 3P mat
When the visitors come,
they should be given a mat.
Hausa is a
predominantly aspectual language so that most TAM paradigms can be used to
refer to past and future events. In (18a), the relative perfective su‑ką 3p‑RP has a past
interpretation in a narrative-like context, while in (18b), the same TAM marker
has a future interpretation in a context where the hearer is given some
instruction concerning a future situation.
When one turns to
the simple temporal clauses, one observes a more restricted number of possible
TAM paradigms. The possible paradigms and their interpretations in simple
temporal clauses are presented in Table 2.
Table 2: TAM paradigms and their
interpretations in simple temporal clauses (with 3rd person plural su‑ and verb fģta go out)
|
|
Past |
Future |
||
|
Single |
Recurrent |
Single |
Recurrent |
|
|
Completive: dą sun fģta |
|
Yes |
Yes |
Yes |
|
Rel. Perfective: dą suką fģta |
Yes |
|
|
|
|
Imperfective: dą sunąa fģtaa |
Yes |
Yes |
|
|
|
Future I: dą zaa sł fģta |
Yes |
|
|
|
Table 2 shows
that simple temporal clauses accept only four TAM paradigms: the completive,
the relative perfective, the regular imperfective, and the future I. In
particular, the simple temporal clauses do not normally take the relative
imperfective, despite their admitting the regular imperfective. One may also
note that Table 2 takes into account both single and multiple occurrence
events, since this has no incidence on the ability of the clause to function
adverbially.
Regarding the
temporal interpretation of the TAM, one notices in Table 2 a general shift
to a past interpretation for most tense/aspect paradigms. For example, of all
admissible TAMs, the completive alone can be used to refer to past or future
events, under the conditions given in the table. The two interpretations of the
completive are illustrated in the following:
(19) Dą sun fģta, sai tą rufč koofąa/
when 3P.CPL go.out then 3FS.SUB close door/
ką shāidaa ma-nģ.
2MS.SUB advise to-1S
As soon as they go out, she
(usually) closes the door/
you should advise me.
In this example, a
completive simple temporal clause can modify a main clause referring to
recurrent past or future situations. The other three TAMs in Table 2 can
only be used to refer to past events. This is illustrated in the following:
(20) a. Dą su-ką fģta, sai ta rufč koofąa/
when 3P-RP go.out then 3FS.RP close door/
*ką shāidaa ma-nģ.
2MS.SUB advise to-1S
Once they exited, she
closed the door/ [once they exit]
please advise me.
b. Dą su-nąa fģtaa, sai ta rufč koofąa/
when 3P-IPV go.out then 3FS.RP close door/
*ką shāidaa ma-nģ.
2MS.SUB advise to-1S
When they were going out,
she closed the door/ [when they
will be going out] please
advise me.
c. Dą zaa
sł fģta, sai ta rufč koofąa/
when FUT 3P go.out then 3FS.RP close door/
*ką shāidaa ma-nģ.
2MS.SUB advise to-1S
When they were about to go
out, she locked the door/
[when they will be about to
go out] please advise me.
As the examples
show, a simple temporal clause with the relative perfective, regular
imperfective and future I is compatible only with a main clause describing
past events. The future I in (20c), for example, describes a
future-in-the-past, i.e., the event in the main clause precedes, and
sometimes cancels, the event in the temporal clause. A probable reason for the
shift to past interpretation observed in Table 2 may be the influence of
the ultimate origin of temporal conjunction dą. In Section 3, it
was suggested that simple temporal clauses developed on the model of temporal dą + Noun/Adverb phrases,
which themselves are based on comitative constructions. It happens that the
comitative use of dą very likely
developed from the ultimate function of dą,
i.e., the existential function (cf.
Abdoulaye 2006). In this context, the anchoring of simple temporal
clauses in the past may be a survival of the existential function, which tend
to describe realized situations (on the long survival or influence of the
original semantics of grammaticalized items, see Bybee and Pagliuca 1987: 117
and Hopper and Traugott 1993: 87‑93). Whatever the validity of this explanation, it is
clear that the shift will be difficult to explain in the framework of the lookącii
deletion analysis. [4]
This paper showed
that in Hausa, one cannot straightforwardly derive simple temporal dą when clauses from temporal lookącin dą time that relative clauses
through deletion of the head word lookącii.
Instead, the paper proposes two different development paths for the temporal
clauses. Temporal relative clauses are grammaticalized forms of ordinary
relative clauses, while simple temporal clauses developed very likely when the
preposition dą (introducing temporal
adverbs, nouns, or verbal nouns) turned into a conjunction introducing finite
temporal clauses. The paper shows that the two types of temporal clauses differ
in their semantics, the list of the TAMs they allow, and the temporal
interpretation of the TAMs. These differences show that simple temporal clauses
are derivationally independent from temporal relative clauses.
Abdoulaye, Mahamane L. (2006). Existential and possessive predications in Hausa. Linguistics 44: 1121-1164.
----- (2008). Perfect and perfective in Hausa. Afrikanistik Online 0009-10-13825. http://www.afrikanistik-online.de/archiv/2008/1382/ (page consulted 25 May 2008).
Adam, Jean-Michel (1994). Le texte narratif: Traité danalyse pragmatique et textuelle. Paris: Editions Nathan.
Bagari, Dauda
Muhammad (1976/87). Hausa subordinate adverbial clauses: Syntax and
semantics. Rabat: Imprimerie El Maarif Al Jedida.
Bearth, Thomas (1993). Satztyp und Situation in einigen Sprachen Westafrikas. In: Möhlig, Brauner and Herrmann Jungraithmayr (eds.), Afrikanistentag, pp.91‑104. Köln: Rüdiger Köppe Verlag.
Bybee,
Joan L. and Pagliuca, William (1987). The evolution of future meaning.
In : Ramat et al. (eds.), Papers from the 7th International Conference
on Historical Linguistics, pp. 108-122. Amsterdam: John Benjamins.
Givón,
Talmy (1990). Syntax: A functional-typological introduction, Vol. 2. Amsterdam: John Benjamins.
Güldemann,
Tom (2003). Grammaticalization. In: Nurse, Derek and Gérard Philippson
(eds.), The Bantu languages, pp.182‑194. London: Routledge.
Heine,
Bernd and Mechthild Reh (1983). Diachronic observations on completive focus
marking in some African languages. In: Heine, Bernd et al. (eds.), Sprache
und Geschichte in Afrika (SUGIA) 5, pp.7‑44. Hamburg:
Helmut Buske.
Himmelmann,
Nikolaus P. (1997). Deiktikon, Artikel, Nominalphrase: Zur Emergenz
syntaktischer Struktur. Tübingen: Max Niemeyer.
Hiskett, Marvyn (1971). The Song of Shaihus Miracles: A Hausa hagiography from Sokoto. African Language Studies 12: 71‑107.
Hopper, Paul. and Elizabeth Closs Traugott (1994). Grammaticalization. Cambridge: Cambridge University Press.
Hyman, Larry M. and John R. Watters (1984). Auxiliary focus. Studies in African Linguistics 15(3): 233‑273.
Jaggar, Philip J. (2001). Hausa. Amsterdam: John Benjamins.
König, Ekkehard (1991). The meaning of focus particles: A comparative perspective. London/ New York: Routledge.
Lehmann, Christian (1984). Der Relativsatz:
Typologie seiner Strukturen, Theorie seiner Funktionen, Kompendium seiner
Grammatik. Tübingen: Narr.
Michaelis,
Laura A. (2006). Tense in English. In: Aarts, Bas and April McMahon (eds.), The
Handbook of English Linguistics, chapitre 10. Oxford: Blackwell.
Moussa-Aghali, Fatimane (2000). Yarintata [My childhood]. Niamey: Editions Albasa.
Newman,
Paul (2000). The Hausa language: An encyclopedic reference grammar. New
Haven: Yale University Press.
Reineke, Brigitte (1998). Des
constructions relatives dans les langues de lAtakora. Cahiers Voltaļques 3: 95‑106.
Schachter,
Paul (1973). Focus and relativization. Language 49:
19-46.
-----
(1985). Parts-of-speech systems. In: Shopen, Timothy (ed.), Language typology and syntactic description:
Vol. 1 Clause structure, pp.3‑61. Cambridge: Cambridge University
Press.
Schuh,
Russell G. (1998). A grammar or Miya.
Berkeley: University of California Press.
Thompson,
Sandra A. and Robert E. Longacre (1985). Adverbial clauses. In: Shopen,
Timothy (ed.), Language typology and
syntactic description: Vol. 2 Complex constructions, pp.171‑234.
Cambridge: Cambridge University Press.
Tuller, Laurice A. (1986). Bijective relations in Universal Grammar and the syntax of Hausa. Unpublished Ph.D. Dissertation, University of California Los Angeles.
van der
Auwera, Johan and Kučanda, Dubravko (1985). Pronoun or conjunction The
Serbo-Croatian invariant relativizer to. Linguistics 23: 917‑962.
Vlach,
Frank (1981). The semantics of the progressive. In: Tedeschi, P. and A.
Zaenen (eds.), Syntax and Semantics,
Volume 14, pp. 415-434. New York: Academic Press, Inc.
Wald, Benji (1987). Cross-clause relations and temporal sequence in narrative and beyond. In: Tomlin, Russell S. (ed.), Coherence and grounding in discourse: Outcome of a symposium, Eugen, Oregon, June 1984, pp.481‑512. Amsterdam: John Benjamins.
Watters,
John, R. (2000). Syntax. In: Heine, Bernd and Derek Nurse (eds.), African
languages: An introduction, pp. 194‑230. Cambridge: Cambridge
University Press.
Wolff, H. Ekkehard (1993). Referenzgrammatik des Hausa. Münster/ Hamburg: LIT.
[1] Hausa
(Chadic) is spoken mainly in Niger and Nigeria. Primary data in this paper are
mostly from the Katsinanci dialect and Standard Hausa. The transcription
follows the Hausa standard orthography with some changes. Long vowels are
represented as double letters, low tone as grave accent, and falling tone as
circumflex accent. High tone is unmarked. Small capitals <b, d,
k> represent
glottalized/laryngealized consonants, and <r>
represents an alveolar trill distinct from a flap [r]. Written <f> is
pronounced [h] (or [hw] before [a]) in Katsinanci and other western
dialects. The abbreviations are: 1, 2, 3 1st, 2nd, 3rd person;
cop. copula; CPL completive; DF definite;
F feminine; FUT future; imp impersonal; IPV imperfective;
M masculine; P plural; RI relative imperfective;
RP relative perfective; S singular; SUB subjunctive.
* Abdou Moumouni University, Niamey, Niger
[2] This pseudo
head deletion process is actually thought by some writers to apply to all adverbial
relative clauses. For example, Reineke (1998: 103) reports that in
Ditammari (and other Gur languages), locative and manner adverbial clauses are
headless relative clauses that however incorporate noun class markers
compatible with, respectively, the lexemes meaning place and manner.
However, in the two Gur languages described by Reineke (Ditammari and Biali),
the head words meaning time are not deleted in temporal relative clauses.
[3] There is
nonetheless a difference between a hybrid temporal baayan dą after that/when illustrated in (16a) and a temporal baayan after clause illustrated in
(16b). The simple baayan clause
allows a more or less extended time between the two events. By contrast, the
hybrid baayan dą clause typically
implies a relatively short time between the two events. Besides koo even and baayan, the conjunction dą
combines with other particles, such as tun
since, sai only, then, to express
a close consecution between events (usually with an implied causal relation).
[4] Beside the
facts reported in this paper for Hausa, the deletion analysis is also
problematic on principled grounds. Indeed, in this analysis, the word lookącii time, which surely keeps some
semantic load, is deleted while the semantically empty relative conjunction dą acquires, after deletion of lookącii, the more substantive meanings
of as soon as, once, when, as, because, etc. This normally violates known
grammaticalization tendencies. In fact, if something should be deleted in a
temporal relative clause introduced by
lookącin dą time that, it would probably be the weaker grammatical
element dą. For example, Schuh
(1998: 272) reports that some temporal clauses in Miya (Chadic) can be
introduced by młkwį day [that] or młku ma day that, i.e., the weak head młkwį midday (a temporal word derived
from młku sun, day) requires no
relative pronoun/conjunction.