Sereer has a moderately large consonant inventory with 32 distinctive consonant phonemes. Strikingly, there are 21 phonemically distinctive oral stops. The consonant inventory makes use of some combinations of parameters that are unusual cross-linguistically, including contrastive voicing in both egressive and ingressive stops. The use of implosives is fairly common in languages of sub-saharan Africa, particularly in the "Macro-Sudan belt" (cf. Güldemann 2010) that spans the non-desert regions and sahel from Senegal to Sudan. The use of voiceless implosives, however, is considerably less common, even within this area (cf. McLaughlin 2005).
Sereer also has multiple uvular phonemes, a rarity for sub-Saharan Africa.
The consonant inventory is shown below. Orthography for a given symbol is indicated in parentheses following a symbol if the orthography differs from the IPA.
|Prenas.||ᵐb (mb)||ⁿd (nd)||ᶮɟ (nj)||ᵑg (ng)||ᶰɢ (nq)|
Sereer has a highly unusual and large stop system, with 21 phonemic stops contrasting across six places of articulation (bilabial, alveolar, palatal, velar, uvular, and glottal). The size of the stops as a component of the consonantal phonological system can be attributed to the fact that Sereer distinguishes five manners of articulation in addition to six places: plain voiceless, plain voiced, prenasalized voiced, voiced implosive, and voiceless implosive. The latter stop manner is especially unusual cross-linguistically, and Sereer has a particularly unlikely system in that there are three voiceless implosives.
All five of these manners are found for the labial, alveolar, and palatal places of articulation. For the velar and uvular places of articulation, there are no ingressive stops; there is also no plain voiced velar stop. This is not entirely unexpected given what the Aerodynamic Voicing Constraint tells us about the antagonistic relationship between obstruency (closure) and voicing (airflow): stops that are articulated back further in the mouth will tend to have fewer manners of articulation that involve passive cavity expansion (e.g. to maintain voicing) simply because there is less cavity space to build up air pressure in. Once a threshold of pressure is reached, airflow ceases, and so does voicing. Sereer's uvulars and velars are entirely consistent with this model, given that the voiced segment present the furthest back (uvular /nq/) has prenasalization as a mechanism for maintaining continuous airflow despite the small cavity size.
Notes on realization:
- All stops with voicing tend to be partially devoiced before pause. This is especially noticeable for modally voiced stops, which are actually aspirated at times in this position (e.g. /b d g/ > [b̥ʰ d̥ʰ ɡ̊ʰ]).
- Prenasalized stops are realized as nasals when preceding another nasal. This effect also occurs gradiently as speech rate increases.
- The implosive stops (both voiceless and voiced) are often realized with a considerable amount of creak on the neighboring vowels. A telltale sign for a voiceless implosive (as opposed to a voiced one) is the impression of a glottal stop plus a voiced implosive, with extra creak on the preceding vowel.
- All implosive stops lack an audible release burst, even in word-final position.
- The uvular stop /q/ is generally accompanied by frication, often pre-frication. It may even be accompanied by some pharyngealization or pharyngeal frication.
- Velar stops are occasionally realized as uvulars when a uvular is present nearby, resulting in a sort of uvular-dominant "dorsal harmony". This is highly variable and is likely due to sporadic processing errors.
Sereer has three voiceless fricative phonemes: /f s χ/. This portion of Sereer Salum's phonemic inventory distinguishes it from Sereer Siin, which has four phonemic fricatives /f s χ h/; Sereer Salum has merged the latter two into the single phoneme /χ/ (see also S and X Alternations).
Notes on realization:
- The uvular fricative varies in the amount of frication it occurs with, and can sometimes approach [h]. This is especially true in fast speech and before front vowels. In contrast, when carefully enunciated it can approach a voiceless uvular trill [ʁ̥] and is often accompanied by a significant amount of pharyngealization.
- /f/ varies between [f] and [ɸ].
There are four nasal phonemes in Sereer at labial, alveolar, palatal, and velar places of articulation: /m n ɲ ŋ/. Their realization is not particularly variable.
Glides, Liquid, Tap/Trill
Sereer has two glide phonemes /w/ and /y/. There is an alveolar trill /r/ that surfaces as a tap [ɾ] in quick speech. There is also an alveolar lateral approximant /l/.
Notes on realization:
- /w/ is realized as front-rounded [ɥ] when preceding the front vowels [i e].
The glottal stop in Sereer is not contrastive word initially, as there are no minimal pairs in this position. ‘Vowel initial’ words are preceded by a glottal stop in careful speech, but it may be lost in fast speech. It is however, contrastive in other positions. For example, it distinguishes past tense forms from present tense forms for verbs: loolaam ‘I laugh(ed)’ vs. lool'aam ‘I laughed (longer ago)’. It is also contrastive word finally: gar ‘come’ vs. ga' ‘see’. Interestingly, Sereer distinguishes sequences of glottal stops across morpheme boundaries. The sequence [ʔʔ] results when a verb ending in /ʔ/ is takes the past tense suffix -/ʔ/-: ga'aam 'I see' vs. ga''aam 'I saw (longer ago)'.
Sereer has a 5 vowel system that is contrastive for length. There are no diphthongs: coda /w y/ may give off this impression, but they can be analyzed as precisely that. There are no restrictions on the appearance of long or short vowels in any environment.
|High||i ii||u uu|
|Mid||e ee||o oo|
Notes on realization:
- Vowels after implosive consonants are sometimes creaky; this, however, is not contrastive.
- /o/ is realized as [ɔ] in closed syllables. See ñaaƴloox [ɳaʄlɔ:x] "feces" vs. laalo [la:lo] "baobab leaf".
- /e/ has a lax variant [ɛ] that also surfaces in closed syllables. See saate [sa:te] "town, village" vs. yeeyet [je:jɛt] "insect".
- Tenseness generally covaries with vowel length; long vowels are more tense than short vowels.
- /a/ is realized somewhat higher than cardinal [a] when short and in a closed syllable. It is particularly susceptible in this position to coarticulation with neighboring consonants, as in the word 'cook' jaw, which is consistently realized as something like [ɟəw].
Sereer disallows vowel hiatus both word internally and at word boundaries. This is resolved either through deletion, in the case of word-external hiatus, or through vowel allomorphy or epenthesis when word-internal.
In cases where VV occurs at a word boundary, V1V2 --> V1. This is evidenced by the deletion of the initial a- indicating differential object marking when an animate object follows a vowel-final verb.
There are two competing hiatus resolutions for word-internal VV sequences: allomorphy and epenthesis. These processes come with a caveat: in many cases, they may in fact be morphologically conditioned and specific to a small subset of inflectional morphology.
VV --> V phonological changes word-internally are often found in interactions between inflectional verbal morphemes. One regular change that occurs is [ao] > [a] as shown by the alternation naf-a-ong > [nafang]. There is also an alternation between [ii] ~[ee] in the negative affixes. The conditioning environment for this difficult to determine (due to the competing hiatus resolution described below) and more research is needed on this.
When a vowel-final verb root attaches to a vowel-initial suffix, either a glide or a glottal stop can be inserted. fi "to do, make" with first person subject marking always surfaces as fi'aam. However, with third person agreement, we see a glottal stop: afiya.
Stress is assigned metrically in Sereer. Stress is assigned to the left-most long vowel in a word; if there are no long vowels, then stress is assigned to the first stem syllable. Thus, stress is often realized on the penultimate syllable of a word.
Due to these stress-assignment rules, many word-initial V syllables are not stressed. Coda consonants are extra-moraic, and thus do not seem to attract stress assignment. It is unclear at this point what the phonological correlates of stress are in Sereer. However, pitch or intensity or a combination of both can serve as a cue for a stressed syllable.
It is suggested that instances of consonant mutation that appear at multiple discontinuous locations are suprasegmental features are applied at the word level. It is unclear how well this analysis is supported, however, and is not necessarily descriptive, but rather theoretical.
Borrowed words show phonemes that are not normally present in Serer, like /y/ in [myyr] 'wall' (Fr. mur), or /ʃ/ in [maʃin] 'machine, device' (Fr. machine). These phonemes are irregularly adapted to Sereer phonology, as has happened on occasion to "wall," which can be pronounced as [miir].