Volume 13 Issue 1 (2015) DOI:10.1349/PS1.1537-0852.A.448 Note: Linguistic Discovery uses Unicode characters to represent phonetic symbols. Please see Optimizing Display for requirements to accurately reproduce this page. A Tasawaq (Northern Songhay, Niger) Text with Grammatical Notes Maarten Kossmann Leiden University 1. Introduction[1] Tasawaq is a Northern Songhay
language spoken in the oasis In-Gall about 100 km west of Agadez in Niger (for
more information, see Bernus & Bernus 1972; Sidibé 2002). The language is
poorly documented,[2] and only little textual material is
available to the scientific community. As far as I know, only two texts have
been published in Tasawaq,[3] neither of them marking tone; in
the first place a relatively long text of oral history edited by Pierre-Francis
Lacroix in Bernus & Bernus (1972:107-114), in the second place a 20-line text
in Rueck & Christiansen (2001). In this article, I present a Tasawaq
story with glossing and comments, recorded in Agadez in October 2003, told by Mrs. Ibrahim, born Nana Mariama Aweïssou,
originary from In-Gall, but then living and working in Agadez. Mrs. Ibrahim
speaks Tasawaq, Hausa and French; at the time of the recording her daily
language was Hausa. Since Lacroix (1971), Tasawaq is normally
considered a mixed language of Songhay and Tuareg, a view that was elaborated by
Robert Nicolaï (e.g. 1990), as well as by Alidou (1988) and Wolff & Alidou
(2001); for more agnostic views, see Kossmann (2007); Souag (2012). In the
text, non-Songhay etymologies have been identified in the glossing line by
means of superscript TU (Tuareg), HA (Hausa), and AR
(Arabic). 2. Notes on Phonological Processes
and Transcription[4] Mrs. Ibrahim’s language has a number
of features that differ from those described in Alidou (1988), Nicolaï (1979;
1979-1984; 1980) and Sidibé (2010a). While in some cases this may be due to
analytical issues, at a number of points the differences clearly reflect
dialectal variation within the language (Sidibé 2010a). Most salient among
these is the existence of pharyngealized consonants in Mrs. Ibrahim’s speech,
whereas the speakers underlying the other sources do not have it (Kossmann 2012). The transcription follows the
surface phonemics of the language, writing neutralizations and assimilations
wherever they lead to differences on the phonemic level, but not writing them
where they lead to allophonic variants. Some of the more important segmental
phonemic and phonetic issues are the following: (a) Short /e/ and /o/ are neutralized into /a/ when not in word- or phrase-final position. Lowering of /o/ may be accompanied by labialization of adjacent velar and uvular consonants
(cf. Kossmann 2012). Examples:
(b) Long vowels only occur in non-final
open syllables. A number of morphological processes lead to the opening of a
closed syllable, accompanied by lengthening (or, seen from a different angle,
show the original long vowel), e.g.
In connected speech, long vowels sometimes
appear in unexpected positions. Some of these are the result of vowel
coalescence (although this mostly leads to a short vowel), others involve CV
stems. The exact conditions of such lengthenings are unclear. (c) In closed syllables with a nasal
coda, the nasal is obligatorily realized as nasalization of the vowel when
followed by a fricative or a glide. In other contexts, there is variation
between nasalization and the presence of a nasal stop, which is not entirely
predictable. Because of this, nasalization (written by superscript <n> following the vowel) and nasal consonants are transcribed
differently. It is very well possible that a more elaborate study would show
that the two are in fact free or idiolectally conditioned variants.
(d) Velar stops are strongly
palatalized in contact with a front vowel /i/, /e/, [æ]~[ɛ], the latter being non-pharyngealized realizations of /a/. The outcome of palatalization is either a palatalized consonant [kj], [gj], or, in the case of /g/, a plain palatal stop [ɟ]. As
palatalization is entirely predictable, it will not be written here. (e) There is a strong tendency to
devoice vowels between voiceless consonants and in final syllables. This makes
it often difficult to hear the vowel, and especially to establish its tone.
Although clearly a phonetic feature, which may be an idiosyncrasy of Mrs. Ibrahim’s
speech, I write the devoicing in the transcription in order to indicate that in
such situations both the tone and the vowel quality are uncertain. In some
contexts, the tone of the devoiced segment can only be determined by its
effects on downdrift. (f) The role of consonantal length
is not entirely clear. Some short grammatical morphemes are frequently
geminated in intervocalic position (e.g. Hǹ
‘genitive’; ní ‘Negative Perfective’).
Vaccillating consonantal length also appears with some other morphemes, but
without a clear conditioning (e.g. qá ~ qqá ‘all’). I write consonantal length wherever I hear it. Tasawaq tone has only received
limited attention, and at many points my notations do not concur with existing
descriptions, e.g. Nicolaï 1980, Alidou 1988, Sidibé 2010a. The tone system
found in Mrs. Ibrahim’s speech has the following properties: (a) There are two tones, High and
Low, and one contour tone, Falling. There is, phonetically, no rising tone,
except sometimes in vowel coalescence. The language has downdrift. (b) Falling tone only occurs on long
vowels, and on closed syllables with a sonorant as their coda, e.g. gáàsù ‘cheese’; hâmnì ‘flour’; sèèrây ‘friend’; àlxâl ‘situation’. There is one case of a Falling tone on a
closed syllable of a different type: yâddà ‘still’. As a result of vowel coalescence, in
connected speech, sometimes Falling tones appear on phonetically short vowels. (c) In isolation, there are no polysyllabic
words with an all-Low tone pattern (differently Nicolaï 1980:248-250). However,
in a number of syntactic contexts, words do appear in an all-Low tone pattern.
Such words have an initial Falling tone in isolation, or, when the syllable
structure does not allow for a Falling tone, they have an initial High tone. As
there are other words which keep their original tone pattern in the same syntactic
contexts, I consider words with variation between all-Low and other patterns to
be underlyingly all-Low. The contexts where the all-Low
pattern appears are the following: -with nouns, when they are followed
by an adjective, a numeral, the plural clitic H-yo, or a postposition. The isolated
form is used with the demonstrative element L-ɣo.
-with nouns, when preceded by a possessor
phrase, e.g.
-with verbs when they are followed
by a direct or indirect object, e.g. with hângwày ‘think of’ and qáɣàm
(< qâɣàm) ‘chew’:
(d) A number of elements take polar
tone, i.e., their tone is the opposite of an adjacent tone. Polar tone is found
on the following elements: *Subject pronouns: -Third person subject pronouns have
a polar tone depending on the following element, e.g.
-The same is true for the marker of
the plural imperative, wa, e.g.
-There is variation between stable
High tone and polar tone with 1S and 2S subject pronouns; this could be a
difference between isolated forms (stable high tone) and clitic forms (polar
tone), e.g.
*Oblique pronouns: -Third person and 1S and 2S direct object
pronouns have polar tone to the element preceding it, e.g.
*The dative postposition -si has polar tone to the element
preceding it,[5] e.g.
Other postpositions have a stable
tone (e.g. gá ‘in’), or the situation is unclear. *The clitics H-yo ‘plural’ and L-ɣo ‘demonstative’, and probably some
other clitics, have polar tone to the element preceding them, e.g.
(e) A number of elements are
preceded by a floating tone, which attaches to the preceding element. The most common
cases of this are: *The plural clitic -Hyo, e.g.
*The demonstrative element -Lɣo, e.g.
*The genitival postposition Hǹ (often geminated in intervocalic
position), e.g.
Due to nasalization, and sometimes
subsequent denasalization, the Low-toned MAN marker m̀ is often mainly realized as a Low tone. Similarly,
the genitival postposition often functions as if it were a floating Falling
tone; in the latter case, however, nasalization is never undone. (f) Due to vowel coalescence or to
the attachment of a floating tone, sometimes an infelicitous tone pattern is
generated. Infelicitous tone patterns are either Rising tones, or Falling tones
on open syllables with short vowels, or Falling tones on closed syllables with
a non-sonorant consonant in the coda. The following tone rule accounts for most
(possibly all) cases: →R and infelicitous F are reduced to H when following
a Low tone, and to L when following a High tone.
3. Grammatical Notes There is relatively little available
on the grammar of Tasawaq. The most comprehensive overview is found in the
unpublished MA Thesis of Alidou (1988), summarized in Wolff & Alidou (2001).
The latter publication focusses on the relationship between elements with a
Songhay etymology and elements with a Tuareg etymology, a focus shared with
Kossmann (2007). A number of more detailed questions have been treated in
Sidibé 2010a, 2010b, Kossmann 2008; 2009; 2010a; 2010b; 2011. As my data are
not always entirely identical to Alidou’s, I think it is useful to provide some
basic notions of Tasawaq morphology below. Genitival constructions,
relativization and adjectives will not be treated, as they were already
analyzed in Kossmann 2009, 2010a, and 2011, respectively. 3.1 Personal Pronouns Tasawaq distinguishes between emphatic
pronouns and clitic pronouns. Emphatic pronouns have their own tone, while many
clitic pronouns have polar tone, i.e. they take the opposite tone of the
adjacent element in the verbal complex. There is no difference between the two
sets in the first and second person plural. The latter pronouns do not cause vowel
lengthening in CVC verb stem, which suggests that they are not cliticized in
any context. In the Imperative, a special marker
for the plural addressee is used.
For the first person subject pronouns, the allomorph ɣay / ɣáy is used when no overt MAN marker follows, while otherwise ɣa / ɣá is used. The clitic forms are also used in combination with
postpositions. The allomorph ɣa (1S) is used with the postposition si ‘to’. 3.2 Nouns There is a major divide between nouns of Songhay and nouns of Tuareg
origin. Nouns of Tuareg origin have lexical (and highly irregular) plurals,
while nouns of Songhay extraction use a NP-final clitic -Hyo (polar yo preceded by a
floating High tone). For details, see Sidibé 2010a; Kossmann 2007, Kossmann 2010b.
Nouns of Tuareg origin denoting human beings allow for gender derivation, e.g.
This is found with only two nouns of non-Tuareg origin:
Otherwise, natural gender is expressed by suppletion, or not expressed
at all, e.g. àlzírày ‘male or female in-law’; ízè ‘son,
daughter’; àssàbí ‘male or female child’ (< Arabic); áàrù ‘man’ – wây ‘woman’; báynà ‘male slave’ – ṭààmú ‘female slave’.
With Tuareg-based nouns, gender is also found differentiating fruits from their
trees, and feminine gender generally expresses language names, e.g.
3.3 Verbs Verb stems in principle do not change. There are, however, a few
processes that apply when the verb is followed by a direct object clitic. In the first place, a number of verbs of the structure CV̀nV́ have
clipped forms (CV́n) when followed by a first or second singular direct or
indirect object clitic, e.g.
In the second place, many verbs have vowel lengthening when followed by
a third singular or plural direct object pronoun. Both these pronouns are
vowel-initial; note however that the effect does not appear with the first and
second person plural pronouns which also start in a vowel. The lengthening of
the syllable sometimes shows underlying vowels obscured by the neutralization
processes applying with short word-internal vowels. All CVC verbs have
lengthening, e.g.
Vowel lengthening is also found with some disyllabic verbs. These
include pluractional derivations from CVC verbs, and verbs derived by means of
the deictic element -nàn.
Vowel lengthening is also found with a small number of underived
disyllabic verbs:
Note that with most underived disyllabic verbs, and with verbs derived
by the suffix -kàt(é), there is no vowel lengthening, e.g.
Mood, Aspect and Negation (MAN) are expressed by particles (some of
which are grammaticalized verbs) that are put immediately before the verb stem:
Alidou (1988:54) has <má> instead of
m̀. This is
probably a case of idiolectal or dialectal variation. Mrs. Ibrahim never has a
full vowel with the subjunctive, and the tone is clearly Low. The element kwáy comes from the verb kwáy ‘to go’, while tí is no doubt related to tê (underlying form) ‘to
come’. In the negative future, the element b is often assimilated to the following k, i.e. sí-k-kwáy-. The main uses of the MANs are as follows: Perfective: punctual events that took place in
the past (for examples, see the text) and states, e.g.
Imperfective: habitual and progressive, e.g.
In addition to this, some stative expressions
use the Imperfective, e.g.
The exact distribution of Perfective and
Imperfective expressions of state is unknown. Subjunctive: In main clauses, the subjunctive
expresses a wish or an order. In subordinate clauses, it is used to express
finality, e.g.
Future I (< ‘go’) and II (< ‘come’) express
the future. There is no obvious difference in use between the two, e.g.
The expressions have become fully
grammaticalized. It is not possible to have a subject marker between kwáy- or tí- and the main verb; moreover
it is possible to combine the future marker kwáy with the verb kwáy ‘to go’, e.g. ɣà b-kwáy-kwáy ‘I shall go’. In addition to these grammaticalized verbs in
MAN constructions, there is one other multi-verb constructions of a similar
type, the intensive marker ṭán (= ‘to be full’). The exact
structure and meaning of this construction need more investigation (see Alidou
1988:51 for more examples). Example:
There are a number of verbal derivations in
Tasawaq: 1. The productive directional derivations kà ~ kàt ~
kàté ‘towards the speaker’ (Ventive)
and -nàn ‘away from the speaker’ (Itive), which
historically derive from verbs (cf. Zarma kàté ‘to bring’; náŋ ‘abandon’), e.g.
The derivational nature (rather than
considering them clitics or adverbs) of the directional suffixes is shown by
two arguments: -directional suffixes precede object clitics.
In the case of -nàn, the suffix undergoes vowel
lengthening when followed by a third person pronoun, just like verbs. Examples:
-directional suffixes can be part of the
verbal base of an adjectival derivation, e.g.
When the directional suffixes are attached to
a CV̀nV́ verb stem with a final i or u, the final stem vowel is clipped before the
suffix., e.g.
One verb shows irregular changes in stem
shape:[8]
2. Pluractional derivation marked by full
reduplication. The examples that were collected concern monosyllabic verbs:
3. The causative derivation -ǹdá.[9] This is the one valency-changing derivation in Tasawaq. In the variety
spoken by Mrs Ibrahim, it is restricted to a lexically determined set of verbs:
Many verbs do not allow for a derivation with -ǹdá, although they would conceptually be proper inputs
for a causative derivation. Thus góódày ‘to heal’ and fàrɣá ‘to be tired’
allow for the causative derivation, but dóórí ‘to be ill’
does not. Different from our data, Alidou (1988:50) describes the causative
derivation as regular and productive. She provides a number of examples, some
of which were not accepted by Mrs. Ibrahim, such as <wándà> ‘to make eat’ and <nínəńdà> ‘to make drink’. Note that -ǹdá cannot be analyzed as the comitative/instrumental preposition ǹdá and the following object as the prepositional
complement. This analysis is impossible, as -ǹdá and the Direct Object can be separated by other elements, which could
not be the case if it were part of a prepositional phrase, cf.
Tuareg has a productive causative derivation, and there are some cases
in Tasawaq where a Tuareg underived verb corresponds to a Tuareg causative
verb, or where a Songhay underived verb corresponds to a Tuareg causative verb.
Different from some other Northern Songhay languages (cf. Christiansen 2010),
this takes place only sporadically, and such cases are best considered lexical
causatives. Examples:
Otherwise, constructions with the verb dán ‘to make’ are used to express a causative relation,
e.g.
A certain number of verbs are labile in their syntax, i.e. they can both
be employed as transitives and as intransitives, in which the element that is
the direct object in the transitive construction functions as the subject in
the intransitive construction, e.g.
3.4 Verbal Nouns There are a number of derivations that make nouns out of verbs. Some of
these also apply to verbal and to nominal bases. 3.4.1 Action Nouns The most generally found derivation is the action noun. With verbs of
Songhay origin, Mrs. Ibrahim’s variety mostly
has zero derivation. In one class of verbs the tone changes, while there are a
few residual cases of suffixation. Verbs borrowed from Tuareg have Tuareg
verbal nouns. The majority of Tasawaq verbs have verbal
nouns identical to the verb. Their nominal nature can only be shown by their
syntactic behavior as the head of a noun phrase, e.g.
Verbal nouns of Low-tone monosyllabic verbs
change their tone to Falling. I have not been able to determine the tone of
verbal nouns of Low-tone monosyllabic verbs which have syllabic shapes where
Falling tone is excluded. Examples:
A small number of Songhay-based verbs (eight
in my corpus) add a suffix to the verb in order to make the corresponding
verbal noun. Sometimes this suffixation is accompanied by other changes. The
suffixes are -yó,[10] -ní and -ànhà (probably from (H)ǹ hà ‘the thing of’):
3.4.2 Derivation of Abstract Nouns by Means of the Suffix tèèré Abstract nouns can be formed by attaching the suffix tèèré to a substantive, an adjective, or a verb stem, e.g.
3.4.3 Derivation of Agent Nouns with the Suffix kway The suffix -kway (polar tone) is used in order to
derive agent nouns. The derivation expresses that the person involved is
(habitually) closely related to, or defineable by the action or the object it
is attached to. Examples:
The basis of derivation is the verbal noun, as shown by forms such as àlámàx-kwáy, which has the verbal noun àlámàx rather than the verb yílmàq. The use of -kway derived nouns is common with
professions, e.g.
In some cases, a borrowing denoting a profession has received the suffix
-kway, while the basic noun was not taken over: téélà-kwáy ‘tailor’ (< Hausa teelà) ; lìkítà-kwáy ‘doctor’ (<
Hausa likità). 3.4.4 Derivation of Nouns by Means of the Prefix àmà- A number of nouns are derived by means of the originally Tuareg prefix àmà-, with changes in the
tonal and segmental structure of the stem. Nouns with the prefix àmà- denote persons with characteristics defined by the verbal stem.
While most nouns with àmà- have a Tuareg origin, derivation from a Songhay stem
is attested in one case:
Nouns with the prefix àmà- typically allow for feminine derivation, as is
usually the case with Tuareg-based nouns referring to persons. 3.5 ‘be’ Verbs Tasawaq has three sets of verbs that can be translated as ‘to be’ in
English: those expressing locality, those expressing attributes and
identification, and those expressing existence. It is not clear to what extent
these verbs can be combined with imperfective MAN morphemes.
The markers of locality, existence, and negative identity function like
normal verbs, e.g.
The positive form of the attributive ‘be’ verb has special syntax.
Different from other verbs, it is placed after its predicate, and always
preceded by a pronoun. Examples:
In the negation, it is possible to have the negated attributive marker
between the subject and the predicate and still the (positive) attributive
marker in final position:
Probably, the negative ‘be’-verb ssí is
historically a contraction of a negative marker sí and the
positive ‘be’-verb sí. In the actual language, ssí functions as a verb on its own, as shown by nominalizations such as yàrdâ n
ssí ‘doubt, lit. the not-being
of belief’. 4. The Text with Annotations The story was told in an
elicitation-like context in Mrs Ibrahim’s house, with only the story-teller and
the researcher present. In spite of the lack of natural context, it was told
with great confidence, and hardly contains any hesitations or false starts. The
story is told in a very lively way and with great humor. A first transcription
of the text was made with the help of Mrs. Ibrahim. The present transcription
is based on careful re-listening of the recording on the basis of this first
transcription. The text presented here is a
well-known story in the region, a version of which appears, for instance, in
Jacques Pucheu’s collection of Nigerien Hausa stories (Pucheu 1982:45ff.). There
is a clear connection to Hausa stories in the name of one of the participants,
the bóóráy tree. Mrs.
Ibrahim explained that she knew this tree only from traditional stories. The
name clearly reflects Hausa ɓauree ‘fig tree’ (Abraham 21962:91, ‘ficus gnaphalocarpa and other
varieties’), which appears in Pucheu’s Hausa version of the same story in the
same role. Hausa influence in the story telling
is also evident from the formulae opening and closing the story, which are both
from the Hausa tradition. In the transcription, // indicates a
major break in intonation. Elements that I could not interpret are put between
square brackets and glossed and translated as [???]; when I could not make a
reasonable guess at the phonetic shape, it was transcribed [xxx]. Elements
between normal brackets are either unexpected elisions, or elements whose
presence is not certain.
gààtán
gààtánkù̥. This is the common
Hausa opening formula ga ta nan ga ta nan ku ‘here it comes, here it comes for
you’ (Ahmad 1997:18). tàžíítìkùm. I have not been able to make out the meaning
of this phrase. It resembles the well-known north-African opening formula ḥažit-kum (‘I told you’) fairly well.
wày káyná. Lit. ‘little woman’; this is the normal
expression for ‘girl’. The isolated form of ‘woman’ is wây. á-ǹ nàànà. The isolated form of ‘mother’ is náànà.
ẓáàẓí. Allegro variant of záɣzí ‘day’, which was the form given in elicitation.
sáy. This discourse marker is borrowed from Hausa. The frequency of its use
may be a feature of personal style. sín. The verb sìní ‘to say’ is often clipped to sìn or sín. The exact conditioning of the choice of the tone in these clipped
variants is not clear. wày káyná
íngì-qáá-s(ì). Doubling of the
NP by an independent pronoun followed by qá is the regular way of expressing totality. The phrase
clearly shows the NP-final position of the dative clitic. táẓàɣàm.
This expresses all kinds of
preparing oneself for a feast: putting on make-up, good clothes, bracelets and
other jewelery. The verb and the identical noun ‘grooming’ look like a loan
from Tuareg, but I have not been able to identify a source.
nààná-yò. From nàànà-´yo.
bóóráy. The tree was only known to the speaker from stories. It represents
Hausa ɓauree ‘fig tree’. bóóráy ǹn ízè. The genitival postposition
is geminated in between (semi)vowels. The expression ‘child of (a tree)’
indicates fruits. The ‘child’ expression is also used for nuts and kernels,
e.g. táynì nn
ízè (< táynì ´ǹ
ízè) ‘date kernel, lit. child of the date’.
gá. The element gá functions as a noun (‘body’), as a postposition (‘in’) and as a subordinator
(‘when’).
ì túúk-à. The form of the verb represents a vowel
lengthening of túk ‘to hide’.
gìsá. From gìsí with coalescence
with the following vowel. ǹ nàànà
ssíí. Relative clause from a
genitival complement (cf. Kossmann 2010a for details).
ì dáán-à. From dán ‘to make’ with vowel lengthening. ì dáán-à à
wá. This double predicate
construction, in which the object of dán is the subject of the second
verb is the productive expression of the causative.
gí. Assimilated variant of gá ‘when’. té. The verb te ‘to come’ seems to be underlyingly Falling, a tone that sometimes surfaces
when the final vowel is lengthened in discourse, e.g. à téè bí ‘he has come yesterday’. hígì-yó. This plurale
tantum form is only used in an adverbial sense, meaning ‘at home’. It is
clearly connected to the noun húgù ‘house’. bàrá. Form with vowel lowering of bàró.
dà ~ dáɣò. While no doubt historically derived from dá(ɣò), ‘place’, the element dà also functions as a postposition. I consider cases
where dà is connected
to the noun by means of the genitival marker Hǹ as
constructions involving the noun ‘place’.
dà. This is the topic marker dà, in this text especially frequent after
personal pronouns. It is not entirely clear whether its tone is always Low, or
rather polar.
nì báɣà. From nì báɣò. Note the polar tone on nì.
dàbdá-yò. From dàbdè-´yo; the form in isolation of the noun is dábdè (< dâbdè).
sèèrááyḁ̀n.
This is a plural of the
originally Songhay word sèèrây. It is one of the few
instances where a Tuareg plural formation is applied to an originally Songhay
word (cf. Kossmann 2007, Sidibé 2010a). ńgì-qí. From íngì-qá with vowel coalescence. ìnhì̥. The exact meaning of the particle is
unknown, but in the text it always occurs in contexts involving movement away
from the speaker. In a number of cases in the text it is used in combination with
a pronoun which denotes the person(s) going away.
tò. From Hausa tô ‘well’.
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||