PHONOLOGY I 3.0 Definition honology is the branch of linguistics which studies the Pose and patterns of speech sounds, i.e. it studies the way sounds form systems and patterns in languages. Phonological sounds knowledge allows us to know what sequence of sounds is permissible or allowed in a language and whether certain sounds are significant or not. Phonology is related to phonetics since the latter serves as its informant. From the knowledge of phonetics (ie. the recognition of the sounds of a particular language) it is possible for a speaker to understand why certain sounds pattern the way they do. In other words, the output of phonetics is the input of phonology. While human beings can produce any sound, not all sounds are used in every language and the arrangement of sounds differ from language to language. Hence, a Yoriba speaker unconsciously knows that [v 8 © 2] are not found in his language. Likewise, Europeans may find it difficult to produce 1 swhile students may have some kind of introduction, however rudimentary, to fy an introduction would some core aspects cf linguistics, it is not likely that suc! ied iouehide include the discipline described in Unit 3. We have therefore decide io ire ee an easier to follow overview of phonology to feed the more tect ah di : (Phonology 11). It is hoped that the many noticeable overlaps are not stylistically + Offensive ----- Ore Yusuf. Scanned with CamScanner= 36 BOSE SOTILOYE the sounds [Kp 6] which are very ee in many African languages and are effortlessly produced by Africans. Because languages have constraints on the occurrence of sounds some languages do not allow consonant clusters while some do not allow word-final consonants, 1.¢. consonants do not end words. Hence even when words are borrowed from languages which allow this sequence, some alterations are made. For example: 1. ENGLISH YORUBA [bleid] puléedi ‘blade’ [sleit] siléeti ‘slate’ [fo:k] £5okb ‘fork’ [kles] kfléasi ‘class’ 3.1 Phonemes A study of the phonology of a particular language also entails the knowledge of the significant sounds (phonemes) of that language. Sounds are said to be significant if they constitute a difference in the meaning of words. In the English words below, variation of the sounds in the medial position brings about a change in meaning: 2. [set] ‘sat’ [sit] ‘sit? [set] ‘set? [so:t] ‘sort’ Vowels [@ 1 © 9;] are significant and are therefore phoneme in the English language because they bring about changes meaning. Scanned with CamScannerPHONOLOGY 37 met Consonants [d t f v 1] are significant in the same way as in the following data 3. [tai] ‘tie’ [die] ‘die’ [fai] ‘fie’ [vai] ‘vie’ [lai] ‘lie’ 3.2, Allophones The variants of phonemes are called allophones. We can illustrate allophones with the phonological representation of the Plural marker in English [-s -z -iz]: 4. [works] ‘walks; 4 [Sops] ‘shopy_-}/( 3 [stri:ts] ‘streets? {gu:dz] ‘goods’ [basiz] “buses’ [busiz] “bushes” [dgadgiz] ‘judges’ : The phoneme /s/ has [s, z, iz] as: different realizations ‘in specific environments as stated below: [s] is realized after all voiceless sounds (except sibilants) [z] is realized after all voiced sounds (except sibilants) [iz] is realized after all sibilants (i.e. sounds similar to /s/) The three sounds are allophones of the same phoneme. We express this with the notation that [s] varies with [z] which also varies with [iz] with a formal statement like: Scanned with CamScanner5, [s~z~ iz] ve some sounds in free variation. It is also possible to ha f These are sounds that. can be interchanged with each other without any change in meaning: AIGBO: 6. miri/ mili ‘water’ * ala / ara ‘madness’ ire / ile ‘to sell’ GBE (also known as Egin in Nigeria) 7. dlo/dro ‘to dream’ alotlo /alotro ‘lizard’ In both Igbo and Gbe, [1] and [r] are freely interchanged with ~ each other without affecting any semantic change. However, in Gbe, the free variation is attested only after Coronal sounds. 3.3 Identification of the Phoneme Three basic principles can be used in establishing significant sounds in a language: 3.3.1 Minimal Pairs Minimal pairs are a set of words which differ in with respect to any one sound, ¢.8- meaning : Scanned with CamScanner >PHONOLOGY 39 So 9 sat bye wall for. fear pat dye well far dear The change in meaning is signaled by minimal change in the units of sounds. Another, example from the Yortba language may make the point clearer: | 10 ra ‘to buy’ 1a ‘to split’ da ‘to pour’ gba ‘to receive’ sa ‘to pick’ ta ‘to sell’ All the sounds which effect a change in the meaning of the above examples are said to have phonemic status. 3.3.2 Complementary Distribution Another way of establishing phoneme is the use of complementary distribution principle. Two sounds are said to be in complementary distribution when one occurs in specific environments where the other is not found as in the data below: KOREAN (Gleason 1955:56) 11 satan ‘division’ Sihap ‘game’ sul ‘wine’ Sixtay ‘dining room’ s& ‘new’ Silsu ‘mistake’ su ‘number’ [8] occurs consistently before [I] but [s] does not. Therefore they are in complementary distribution and belong to the same phoneme. In essence, [s] and [8] are allophones of the same phoneme in this language. Here are other data to illustrate the point further: Scanned with CamScanner40 BOSE SOTILOYE TOJOLABAL (Mexico) (Gleason 1955:57) 12 tsitam ‘pig’ makton ‘patch’ Ritat! (a kind of plant) mut" ‘chicken’ inat” ‘seed’ tinan ‘upside down’ It is noticed that both [f] and [1] are in complementary distribution. While [t] is found in initial and medial Positions, the occurrence of the aspirated ital is restricted to the word. final position. 3.3.3. Analogous Environment If minimal pairs are not established and sounds are not in complementary distribution, th- »rinciple of analogous environment can be used to ide: uy phonemes. Sounds which occur in similar or analogous :nvironments are said to be different phonemes. Let us have a look at the plain {p] and its aspirate [p"] in the following data: KOREAN: 13 pal ‘leg’ pal - ‘door’ pul ‘fire’ prado “wave” iptida “being dressed’ sinpu ‘bride’ cop'a ‘nephew’ stlopk ‘businessman’ Scanned with CamScannerPHONOLOGY 41 They both occur in initial, and medial positions. In > is. In the vein, /d/ and /t/ are different phonemes as the data below ats “evident: (HYPOTHETICAL LANGUAGE) 14 lat ‘moon’ ” kutle * ‘crystal’ dandu ‘beans’ tede ‘some’ pado ‘size’ 3.4 Phonological Processes Native speakers of languages internalize certain rules and these are applied when they speak. These rules are very regular with only a few exceptions and they come about as a result of phonological processes such as ‘assimilation, deletion, insertion, and metathesis which take place in such languages. 3.4.1 Assimilation This is a process whereby contiguous segments influence each other by becoming more alike. In other words segments do adapt to their environments. Note the negative marker in English: 15 possible impossible [imposebl] jis imbalance [imbeelens] decision indecision [indisisnj tolerance _ intolerance ae correct incorrect [inke: Scanned with CamScanneryy a2 BOsk SOTLOYE ilabi b] which are bilabial bilabial nasal occurs before [p, ¢ re " 2 alveolar nasal occurs before [t, d] which are alveolars, [p] a velar nasal occurs before [k], a velar stop. To make a general statement about the manifestation of the nasals in the negative words, we would say that nasals assimilate to the place of articulation of the following stop. The same process occurs in the Yortba language involving the continuous aspect marker: YORUBA 16 [ba] ‘hide’: [mba] “is hiding’ [£5] “break? [mf] is breaking’ {sin} ‘bum’ [isi] ‘is burning’ [ko] ‘write? [nko] ‘is writing’ [gi] ‘climb? [pga] ‘is climbing’ The nasal Consonant adapts to its environment by becoming — like the consonant it Precedes, hence it is [m] before ials, {n] before alveolars and [y] before velars. This is a a, of consonants taking on features of other consonants. a is = to have assimilatory process involving a vowel os oe atures ts conscnant and vice-versa. Vowels cal certain features thi i vicinity, from other vowels within their Assimilatory processes can naturally be explained. When am speak they do not pronounce sounds in isolation but le producing one sound, the Production of the next sound is Scanned with CamScannerPHONOLOGY 43 — already anticipated. Therefore sounds tend to copy the features of contiguous sounds to them. 3.4.2 Deletion Some sound segments get removed in normal speech. Vowels as well as consonants undergo this process: [Link] Vowel Deletion ~ YORUBA: 17. ko + we write . book mu + ‘emi drink — palm wine f + ilé break —_ house GBE (Soremekun 1986) 18. ava + oxi dog bone ovi + agbasa child gba + ox build - house kawé i ‘write book’ , memii \ ‘drink, palm wine’ f51é “break into a house’ > (avixi ‘dog’s bone’ — dvigbasa ‘child’s body’ gbax ‘to build a house’ Scanned with CamScanner44 BOSE s6Tmo6yé [Link] Consonant Deletion - Y ORUBA 19. egingyin > etingin > eégin “masquerad.» dtits > dite > 06t6 ‘truth’ dwiird > dard 3 d6rd ‘morning’ The deletion of consonants may lead further to assinilay Process as observed in the eximples in (19), above. ~ 3.4.3 Insertion A language may add a segment to break an Unalloweg cluster or make a foreign word conform to the Phonology of the host language. This concept is also known as epenthesis, Fo, example, English loan words into Yordb4 undergo the said epenthesis: 20° ENGLISH YORUBA [kles] [kiléasi] ‘class? [teibl] [tébura] ‘table’ [baibl] [bibéh] ‘bible’ [sku:]] [sikwdla] ‘school’ Yoruba inserts a vowel in-between two consonants and adds a vowel at the end of every word to conform with the preferred syllable structure of the language. 3.4.4 Metathesis This is a process whereby segments interchange positions It is not as common as processes like deletion and assimilation: Scanned with CamScannerPHONOLOGY . 45 ee en OOo 21 esk [eks] ‘ask’ eesterisk esteriks ‘asterisk’. sctifiket scfitiket ‘certificate’ In longer utterances, such processes may come out as a ‘slip of the tongue’ as in: 22. Let me sew you your sheet instead of Let me show you your seat. 3.4.5 Summary In conclusion, phonology is the field which studies the patterning of sounds in languages. One of the basic pre- occupations of the phonologist is the identification of the status of sounds in languages (namely, stating whether they are . phonemes or allophones and the conditions under which they vary; free or conditioned). The various phonological processes which are attested in language are also illustrated. Recommended Readings. Gleason, H.A. (1955) A Workbook in Descriptive Linguistics. NY: Holt, Rinehart and Winston. Hyman, L. (1976) Phonology: Theory and Analysis. NY: Holt, Rinehart and Winston. Schane, S. (1973). Generative Phonology. Englewood Cliffs, New Jersey: Prentice Hall, Inc. Scanned with CamScanner46 BOSE SOTILOYE Sorémékin, B.S. (1986) Two Approaches to Aldigy Phonology: Segmental and Autosegmental. University Torin .M.A thesis. ot Scanned with CamScannerob PHONOLOGY I 3.0 Introduction hononology (or phonemics) is concerned with the Prsitstc of contrastive sounds in a language. Each specific language determines which sound is contrastive within its system. For instance, in the English words /keet/ and /ozt/ (‘cat’ and ‘bat’, respectively), the sounds /k/ and /b/ contrast in that they both occur in the same environment (at the beginning of a word and before the same kind of vowel) and they are the sole determinants for the contrast in meaning. An added evidence is the fact that if one is. substituted for each other, there will be a change in meaning. Sounds which occur contrastively like this are called phonemes. 3.1 Phonology, Phonetics and Orthography 3.1.1 Relationship between Phonology and Phonetics The relationship between Phonology and phonetics can be compared (borrowing an analogy from Pike (1947)) to that of the farmer and the cook; the farmer (phonetics) plants, harvests and the cook (phonemics) cooks it. Phonetics is concerned with how a sound is produced, where it is produced, What kind of sounds are being produced, how the flow of airstream is disturbed, how one sound differs from another etc. without reference to the structure of speech patterns. Fhonemics Scanned with CamScannerye p (4g FRANCIS OYEBADE ee d with the position in a y, the other hand, is concerne t ~ where e ad oceurs (initially. ae or finally, that is the distribution of the sound) oe distributional relationship with other phonetically similar so a - a The relationship between phone! Phonemies cay be illustrated with the following words from Yoruba: 1 ki “to greet” : k6 ‘to gather together’ ké ‘to pamper’ ka todie The four instances of k in the examples are not phonetically identical. If you allow yourself to be deliberately conscious of how you produce the sounds you will find out that as you move from ki to ki, the k soundis gradually being produced farther and further towards the back of the mouth. ‘Thus, as far as phonetics is concerned, the four k’s are different sounds: phonemics, on the other hand will group all the four of them together into one phoneme since they are not contrastive. The most frontal k occurs with i and the most retracted k only occurs with u; if it were possible to interchange‘ the two k’s, it will make no difference to the meaning of the word (unlike the interchange of the first sounds in these two words [g®] ‘to be foolish’ and [k5] ‘to refuse’ which will lead to a difference in meaning in Yoruba). 3.1.2 Phonology and Orthography In the preceding section, the infinite possibility of producing different sounds was alluded to which can be studied in phonetics. However, the optimal orthographic (that is Scanned with CamScannerPHONOLogy 49 writing) system is the one which makes use ofa limited number of symbols (representing sounds) to replicate speech on a different medium. English, for instance, makes use of 26 symbols, Yorba uses 25 symbols (Bamgbose i965), Hausa orthography has 26 symbols (Kraft and Green 1973) ete, Of course there are languages such as Chinese which use over a thousand symbols but they are not as €asy to learn as English or Hausa. Phonemics makes it Possible for us to reduce the inventory of phonetic sounds of a language to a manageable proportion for an orthographic system. Phonemics ensures that only contrastive sounds are distinguished for Tepresentation by orthographic symbols. Furthermore, all contrastive sounds are provided with orthographic symbols such that no. sound is under-diferentiated orthographically. The “more Tigorous a phonemic analysis is, the more economical and simple the orthographic system. : ~The possibility of reducing the infinitely different types of sounds to a small number has been found to be significant for radio, telephone and other telecommunication technology. It has been pointed out that it is useful for orthography; it is also useful for spelling reforms and as aid to the teaching of foreign languages in the shape of contrastive analysis. 3.2° Principles of Phonemic Analysis 3.2.1 Test for the Phoneme ot The phoneme has been defined in various! ways but for Our purpose, we will adopt Jones (1967) definition: ‘ “A phoneme is a family of sounds in a given ee Which are related in character and are used in such «WAY that no member ever occurs in a word in the same context as any other member”. Scanned with CamScannereV BRAINUID UIEDAME This definition can be illustrated using the following ae Dl 2. EFIK (Chumbow Lecture Notes, 1977) kané ‘step over!’ sak° ‘laugh at!’ beye “belch? ka ‘deny’ wak° ‘abound/ multiply’ tidép ké ‘T am not buying’ The sounds [k]. [y] and k°] form a family of sounds in Ek (a they are phonetically related to’ each other being all Velar sounds) with [k] occurring at initial (and medial after consonant) position, [y] at intervocalic (that is between vowels) and medial position and [k°] at the final position. Observe that none of the sounds in this family can occur in the same Phonetic context (that is the same environment) as the others, In such a circumstance, as it is illustrated in (2) above, one of the variant forms (usually the ‘one that has the widest distribution) is chosen as the basic Phoneme while all of the others are classified as allophones of the basic phoneme. Thus an allophone is a sound or subclass of sound which on complementary distribution with another sound (the concept ® complementary distribution will be discussed later, a Sotiloye, _in a chapter Preceding this). The alloph Fy relationship within this family of sound can bei formalized this way: Scanned with CamScannerPHONOLOGY — 51 3) 3, tk! > 4 [7] [ey The first test for phoneme is a minimal pair. A pair of words which not only differ in one segment but also differ in meaning (in such a way that the segmental difference may be reasonably held accountable for the. difference in meaning) is called a minimal pair. An example of a minimal pair from Ekparabong is [wém] ‘beer’ and [sém] ‘hide’. This proves that fw/ and /s/ are phonemes. Segmental difference in a minimal pair. may occur at any point in the distributional scale of a word. In English, pill and bill form a minimal pair with different segments initially, pit and pat with different segments medially and pat versus pan also form a minimal pair with different segments word-finally. The illustration above shows that /p/, /b/, /a/, /V, /t/ and /n/ are phonemes of English. Notice that a minimal pair requires that the pair of sounds that differ in the words must occur at the same structural position. The second test for the phoneme is near minimal pairs. In this case, there may be more’ than one difference but the environment is the same. For example [larat] ‘king’ , [arad] ‘queen’ indicate that /t/ and /d/ are discrete phonemes even though the two words do not form a minimal pair. Another pair of examples of the near minimal pair are [mbon] ‘children’ an: [thpdp] which show that /n/ and /y/ are discrete phonemes as are /b/ and /p/. The third test is occurrence in analogous environment. If two sounds occur at identical positions’ before the same type of ' sounds or after the same type of sounds, they are said to occur Scanned with CamScanner52 FRANCIS OYEBADE in analogous environment. The pair of sons sap / pat an a a dog can be used to illustrate this pene and /p/ car analogous environment because they ee Word-initiaty and word-finally and are preceded or fo lowed by the a vowel. The same can be said of /g/ and /d/ in the DEXt set of examples. Further examples can be illustrated with this Set ig data: 4 1 pada’ ‘tocryout? 8. z4ky “waterfalp 2 tazi ‘mustache’ 9 taka ‘moss’ 3 sapa ‘smoke’ 10 vag ‘sixteen’ 4 dafa . ‘baby’ ll fava ‘silly’ 5 bat& ‘choking’ 12 24b4. ‘to Smell’ 6 sagé ‘shoulder’ 13. kafa “procession? 7 baba ‘witch’ 14. taka ‘one-eyed’ Let us assume that we are concemed with Phonemic status of [t] and [d]. Both conson: medially (numbers -1 and 5 for example); identifying the ants can occur both of them can ‘inet phonemes in this language The opposite of Occurrence in analogous environment is occurre! Cccur), they are said to be in We go back to the Efik data in (2), it will be observed that. [k] and fy] occur in mutually Scanned with CamScannerPHONOLOGY 53 exclusive environments; while [k] oc CUS Word-initi never intervocalically), [y] occurs itervocaicany ta (and word-initially). This indicates that (k] and [y ws complementary distribution with each other. It is Possible eh pair of sounds to occur in the same Position but still be in complementary distribution: 5. (Adapted from Chumbow: Lecture Notes 1977) 1 bagu ‘name? = 7 pimum ‘shirt? 2 nimad ‘car’ 8 déigin war’ 3 gidum ‘horse’ 9. danib table’ 4 magu chair’ 10 nunag ‘eraser’ 3 nidum ‘coat? 1 bigu ‘kitchen’ 6 dadzin. ‘chalk’. 12 gugu “food? The two consonants [d] and [dz] occur in mutually exclusive environments. Even though both of them can occur word- medially (3, 6) and word-initially (8 and 9), [d] never occurs before a high front vowel [i] while [dz] occurs only before the high front vowel [i]. 3.2.2 Iustrative Analysis (Chumbow, Lecture Notes 1977, Adapted) 1. \bagu ‘name’ 7 nimum ni 2 nimad = ‘car’ 8 daigin ae 3 gidum = ‘horse’ 9 danib ‘eraser’ 4° magu chair’ 10 nunag ‘kitchen’ 5 nidum ‘coat? = 11g ‘food’ 6. dadgin ‘chalk’ 12 Bus Scanned with CamScanner54 FRANCIS OYEBADE ' a. MAKE A CHART CONSONANTS Low ‘“ b. LIST THE SUSPICIOUS PAIR OF SEGMENTS: bm dng) d@n mp py d: dé y Scanned with CamScannera PHONOLOGY 55 ¢, LOOK FOR MINIMAL PAIRS: [bam] bagu ‘name’ = magu ‘chair’ (therefore /b/, /ma/ are separate phonemes) [d: 9] pidum ‘coat’ : yinum ‘shirt’ (therefore /d/ and /y/ are separate phonemes) [g: 9] didum ‘horse’: Qidum ‘coat’ (therefore /g/ and /y/ are separate phonemes) d. CHECK FOR OCCURRENCE IN ANALOGOUS ENVIRONMENTS: [n: dz] nimad ‘car’: nidum ‘coat’ (occurrence word-initially before [i]) [m:n] nimad ‘car’: nunag ‘eraser’ (occurrence word-medially before [a]) [n: p]_nimad ‘car’: nidum ‘coat’ (occurrence word-initially before [i]) therefore, [n] and [dz],-[m] and [n], [0] and [p] do not fall into the same family of phonemes. ©. CHECK FOR OCCURRENCE IN MUTUALLY EXCLUSIVE ENVIRONMENTS. [n: p] nimad ‘car’: pinum ‘shirt? danib ‘babble’: nunag “eraser wis» Scanned with CamScanneryr 56 FRANCIS OYEBADE [i] while [n] occurs anywhere else; therefo, te ({n] occurs before f the same phoneme with [n} an e [n] and [n] are allophones © basic alternant) [d:dz] nidum ‘coat’: dadzin ‘chalk’ da diin ‘chalk’: dzigin ‘war’ gidum ‘horse’ ({dz] occurs only be! Therefore they are allophones © the basic alternant). fore [i] while {d] occurs anywhere else f the same, phoneme with [d)as This language has six consonant phonemes and two of these phonemes (/d/ and /n/) have allophones the occurrence of which is predictable, f£ DISTRIBUTIONAL STATEMENT Tol lar {a before a [dl elsewhere lef /nol (Lp) before i] ] inl | elsewhere J ly! Scanned with CamScanner——_—— Sinology _ si VOWELS: iu bigu ‘kitchen’: bugu ‘food’ (Minimal pairs. Therefore /i/ and/u/ are separate phonemes) The language has three vowels /i, u, a/ PRACTICE EXERCISES 1; What is the phonemic status of the Plain, and lai consonants in the data below?: ae ASANTE(Kwa, Ghana, Chumbow: Lecture notes 1977) caca © ‘straw mattress’ dc%a ‘he cuts’ dj4 ‘he loads’ aj"4 ‘he carves’ oné ‘he finds’ ». n’a ‘snail’ 38é-. ‘he puts on’ 38"€>) the looks at’ "2, vlWhat is the phonemic status of the labialized and palatalized consonants in Nupe? NUPE:(Kwa, Nigerian, Chumbow & Ejimatwa 1983) agi ‘child’ eg’u ‘mud’ nang’i’ ‘goat’ eg”o “grass ege ‘wine’ g*o grind’ age. ‘tear’ jek”o ‘door ‘cow nagk"o | °° Scanned with CamScannerr 58 FRANCIS OYEBADE 3.3 Alternation When two morphemes are combined, it may hap two phonetically related sounds occur in Predictably ea exclusive environments. The identification and Classifica these alternants is not, strictly speaking, within the phonemics; it is studied within the branch of Phonology c i morpho-phonemics. We shall discuss this Presently, - alleg Sounds may also alternate with each other With; same morpheme (that is minimal - meaningfy Unit grammatical analysis). The alternation may be Predictable of not. Where two sounds alternate such that they cay y interchangeably used without bringing about a change in the meaning of the word, they are said to be in free variation. WW. can illustrate this with the following data: 5. (Adapted from University of Ibadan LIN 152 Practice Sheet, 1980) 1. pat ‘father’ 9. mbat ‘my father 2. tap ‘water’ 10. mdap ‘my water 3. kak, ‘belt’ ll. ygak ‘my belt’ 4. rat ‘milk’ 12. nlat ‘my mil’ 5. kap ‘hat’ 13 pata ‘the father’ 6. lat “milk? 14. tapa’ ‘the water’ 7. (sae ‘rope’ 15. kaka ‘the belt’ 8. tal rope 16. tara ‘the rope’ The sounds [1] and [r] are in free variation since a substitution of one for the other does not change meaning (for example [rat]: [lat] ‘milk’ and [tar]:[tal] ‘rope’). When two sounds occur in free variation, they are said to be allophones of the same phoneme. Scanned with CamScannerPHONOLOGY 59 Two sounds may also occur in conditioned variation When two sounds occur in mutually exclusive environments such that the phonetic shape of one (or the two) of them can be determined by the surrounding environment, they are said to be in conditioned variation. Data from Swedish can be used to jllustrate this point: 6. (Chumbow: Lecture Notes, 1977): 1. mo: ‘may’ 5. smo: ‘snail 2. no:l needle 6. sgo:] ‘stingy’ 3. ny:sa ‘sneeze’ 7. fgy:sa ‘sniff? 4. lo:sa ‘lock’ 8. flo:sa ‘paint’ (Note: m, ] 9 are all voiceless sounds) The nasal consonants and lateral liquid which are voiced occur lementary distribution with the voiceless ones. The in comp! s of their voiced voice-less ones occur as conditioned variant: counterparts because they occur after voiceless consonants (s, f). In other words, the initial voiceless consonants have conditioned the nasals and the lateral to be voiceless. At the level of morphophonemics, alternations may be phonologically conditioned or morphologically conditioned. 3.3.1 Phonologically conditioned Alternants t is a variant yh (thal When a morpheme has an allomorph (| a) which is form of a minimal meaningful unit of ae cours, the . . sy which i s, dependent on the phonemic environment in aiternatit allomorph is said to be a phonologically conditioned Let us illustrate this with data from English: Scanned with CamScanner_- -_ ENGLISH: 1. kaps ‘cups’ 7 boksiz ‘boxes’ 2. bets ‘bats’ 8 foksiz: “foxes’ 3. larks ‘larks? 9 clifs “cliffs 4, kabz ‘cubs’ 10. keivz ‘caves? ; “beds? . fisiz “fishes 5. bedz ‘beds 1 = fishes 6. kigz ‘kings’ 12. dzadziz “judges? The English plural morpheme in the data has three allomorphs. {-s, -z. -Iz}- Each of these allomorphs depends: on the phonemic environment in which it occurs: {-S} is used in the phonemic environment ofa voiceless consonant, {-z} is used in the phonemic environment of a voiced consonant while {-iz} occurs in the phonemic environment of the sibilants ((z}, [5], [dz], or [8])- Phonologically _ conditioned alternations may be automatic or non-automatic. Automatic alternations occur in every morpheme that has the same phonological form in a given phonemic environment. For instance, [bie] ‘bear’, [hie] ‘hear’ [sie] ‘sear’ with the former set occurring in isolation (and also before consonant-initial morphemes) while the latter occuts before: vowel-initial morphemes (e.g. hearing, bearing searing etc.). This alternation occurs automatically in the phonemic’ environments mentioned above. Non-automati alternation, on the other hand, occurs when a phonological conditioned:alternation is not always predictable. In English ® morpheme which ends in [k] usually has an allomorph whe? the [k] is substituted by [s] before [i] (keeOelik ~ kebelisize® ‘catholic ~ catholicism’ [plestik ~ plestisiti) ‘plas : plasticity’ [kritik ~ kritisizm] ‘critic ~ criticism’ ete.) howert it is not all cases of [s] before [i] that are derived from we Scanned with CamScanner———| PHONOLoGy 6 instances of [k] do not alternate before [i] (e.g. blok ~ dlokin} “block ~ blocking’)... stor} 3.3.2. Morphologically conditioned Alternation - Some morphemes have alternants which. are not, strictly speaking, dependent on Phonological Context but rather their occurrence is determined by.. morphemic. or grammatical information. Such alternations are. morphologically conditioned. In English, morphemes such as. hoof, wife and knife take the voiced plural suffix: (waif ..~ waivz, naif ~, naivz, hu:f >-huvz), These morphemes have two allomorphs each: {waif} ~ {waiv}, {naif} ~ {naiv} and {hu:f} ~ {hu:y}. However,’ the choice of one or the other allomorph »‘is. “not determined. by the Phonological environment alone but by the grammatical context of singular/plural: The Point becomes clearer when we compare the plural ‘forms of the mo: rphemes used as illustration with the genitival counterparts of the: ‘Se morphemes: 8. SINGULAR PLURAL, GENITIVE waif waivz waifs hu:f hu:vz, hu:fs naif naivz, naifs NOTE: The genitive forms can be used in such contexts aa ‘my wife’s brother; ‘the cow-hoof’s distinctive mark’, the knife’s edge’, etc.), | Even though the phonological context for a and | Senitive may be argued to be similar, there is no Sar Phonological reason why the plural forms should requis : alternative which ends in a Voiced fricative while’such a chang Scanned with CamScanner62 FRANCIS OYEBADE does not occur with the genitive forms. The morpholo conditioned nature of these allomorphic alternations js emphasized by the fact that there are examples of morph ae end in the voiceless labio-dental fricative but which not have a different allomorphic alternant when pluraliseq ‘ ~ Gifs ‘chief ~ chiefs’). for ically example ti:f 3.4 Procedure of Phonemic Analysis The procedure of phonemic analysis has been illustrate earlier in section two. We shall look at another set of data iy this-section to further illustrate this procedure. 1. pat “father” 9. mbat ‘my father’ 2. tap ‘water’ 10. mdap ‘my water 3. kak “belt” 11. ggak ‘my belt’ 4. rat ‘milk’ 12. nlat ‘my mil’ 5. kap ‘hat? 13. pata ‘the father’ 6. lat ‘milk’ 14. tapa — ‘the water’ 7s, tar ‘rope’ 15. kaka ‘the belt’ 8. tal ‘rope 16. tara ‘the rope’ A. PRODUCE A WORK CHART. The first step in analyzing phonetic data is to produce 3 work chart which will make it easy to identify phonetically related sounds: al Scanned with CamScannerPHONOLocy B. ARRAY THE SUSPECT PAIRS. The next step is to group'the sounds into Suspect pairs (or suspect sets): Suspect Pairs: p:b; t:d; keg; mm; n:y3) my; Lr C. CHECK FOR MINIMAL PAIRS: [k:p] kak ‘belt’, kap ‘hat’ this indicates that /k/'and pi are phonemes, [tr] pat’‘father’ rat ‘milk’ This shows that /t/ and /r/ are phonemes: D. CHECK FOR ENVIRONMENT OF OCCURRENCE: [p:b] © pat ‘father’; tap ‘water’; mbat “my father’ [td] [p] occurs ‘initially, finally while’ [b], occurs: medially after nasal; therefore the two sounds are allophones of the same phoneme’’since they occur ‘in mutually exclusive ‘environments. Since’ [p] has the wider distribution, it can be considered as the basic alternant. tap ‘water’, pat ‘father, ndap ‘my water’ [t] occurs initially and finally while [d] occurs inedialy after nasal; therefore the two sounds are allophones ©: the same phoneme ‘because ‘they are in son fe distribution. Since [t] has the wider distribution, it “be chosen as the basic alternant. Scanned with CamScanner64) FRANCIS OYEBADE * ilk’. ‘my water’, mbat ‘my father’, mel railed ore so does [m]. However, [n] : only followed by alveolar sounds while [m] can only be followed by bilabial sounds. Therefore, they occur in. complementary. distribution and by that token, they represent allophones of one phoneme since they both have alternants. We decide to choose /n/. n:y: nlat ‘my milk’, ndap ‘my water’ ngak ‘my belt”. [n] occurs word-initially and so does [p]. However,, [n] occurs only before alveolar sounds while [p] occurs only before velar sounds. Therefore they occur in mutually exclusive environments [Link]. thus members of the same phoneme. Since they both have the same distribution, any one of them can be chosen as the basic alternant. However, because we have already chosen /n/ as a phoneme for. this language; we must .be Consistent and retain [n]/as the basic alternant here also. [m:p] | mbat ‘my father’; nak ‘my. belt? [m] and [py]. both occur word-initially,.. but [m] only occurs before a_ bilabial stop . while ..[n]. only:, ovcurs before a velar Stop. Since they are in complementary distribution, they are by, that token members of the saine Phoneme. However none of them can be chosen as the basic alternant because we have earlier identified both of them as allophones and the biuniqueness condition (once a phoneme, always.a Phoneme) insists,:that. they must still remain as allophones, Therefore,. in this case too, the basic alternant is /n/. Scanned with CamScanner$$$ miororocy. 6. ts it ‘milk’, rat ‘milk’, tar ‘rope’, nlat ‘my rope’, tara ‘the rope” {i and [r] are in free variation, the fi allophones of the same phoneme. They baie mutually exclusive ‘environments in some ee “ay” occurs after a nasal Stop while . [r] a intervocalically, which further confirms ‘their allophonic status. Both of them have equal distribution which gives us the opportunity to choose any one as the basic alternant; however, since we have chosen /r/ earlier as “the phoneme, We shall retain it as the’ basic alternant. E. DISTRIBUTIONAL STATEMENT. **’ : Phonemes and their allophones in this language p> [b] word - medially after a nasal stop | [pl elsewhere , [d] medially after a nasal-stop [t}elsewhere © {[g] word- medially after a nasal stop** [kJ elsewhere : [m] word - initially before bilabialsounds Inj —> 4» | word - initially before velar sounds [p] elsewhere Scanned with CamScanner66°" FRANCIS'OYEBADE “<< ffl] word ~initially aftera nasal ” Vee} word - initially, finally and intervocatically Note that [1] and [r] are in Sree variation.” ; F. Write the phonetic data out phonemically, This enta .. Substituting all allophones w ith their’ identizied Phonetic forms... : . : Phonemic Data 1. /pat/__ ‘father? 9:)_. /npat/ one My, father 2. fap! ‘water’ 10. /ntap/ “my water 3. Mkak/ “belt” rll.» /nkak/ ‘my belt? 4. frat ‘milk? \ ey 12... /rat/ ; ‘my milk’ 5. Mkap/ ‘hat? 13 /pata/ “the father’ 6. Nath ‘milk? 14. Aapa/ “the water’ 7. far! ‘rope? 15. /kaka/ “the belt’ 8. Ma ‘rope 16. ‘/tara/ ‘the rope’ NOTE: The data just analyzed may’ be Correctly argued to be Corpus for two different branches; Phonemics (1-6 of Phonetic data) and morphophi onemics (7-16 of phonetic data). However, this argum lent does. not negate the Procedure outlined in this section. , : Practice Exercise oo 1 [aBana] ‘Havana’ ig [dutaz] ‘to endure! 2 [bala] ‘ball’ Th [gaat] ‘toleam’ - 3 [baya] ‘rope’ ® [gato] ae 4 [beso] ‘kiss? ® [gola] eee; 5 [boda] ‘wedding” 21 [posaty ee 6 [buto] ‘burro’ 2. [asa] ia Scanned with CamScannerPHONOLOGY: 67 : ees) ‘to OW 5 devel] © “tO0we [nada] ‘nothing’ i Monde] “where” 2B [nudo] te iL [peo] % [peto] “dog? 2 (pipal ‘pipe TZ. [ponderoso] ‘heavy’ 3 pongo]. ‘Tey, [styato)) ign 4 [tengo] ‘Thave’ —B_ [todo] D>: Sie 15 [taBako] “tobacco” 2D [upa] ‘grape’ Note: [] is @ trill while [#]'is a flap] 15 Syllable Structure and Phonotactics. ‘ On the process of analyzing a language, the axidlyst often comes into contact’ with what are called : suspicious segments. These are complex segments which include’ affricates ({8], [42], [ts], [dz], etc.), sounds with secondary articulations ((&"), [¢)). long consonants and long vowels [g:], [f:], [0:1 fe:] etc.), and aspirated consonants, ({p"],and [e). In determining how to interpret them (either as single sounds or as a sequence of two sounds), two principles are taken into consideration: i, Principle of economy of syllable structure ii’ Principle of economy of phonemes. The first principle constrains the analyst from proliferating syllable structures while the latter constrains him from proliferating phonemes. The application of the two principles a be illustrated with the two sets of data to be introduced Slow: ‘ Scanned with CamScanner68 FRANCIS OYEBADE EKPARABONG (Chumbow: Lecture Notes 1977) lL. KS ‘catch 8 ol “eat? 2. 15 ‘see’ 9% té ‘find! 3. b6 ‘divine’ 10. na ‘rain’ 4.5. sf ‘arrive’ 11. 86... ‘stop’ 5. fi ‘blow’ 12. 185. ‘sel? 6. né — ‘defecate’ 13. 25 ‘sell 7... _kw6 ‘die’ The clear-cut syllable structure of Ekparabon; in the form of 1-10. This syllable structure i 11-13 may be analyzed as a sequence of consonants followed by a vowel. If the former Position is adopted, it will introduce a syllable structure CCV , which is different from the clear-cut syllable structure of Ekparabong. This will not be in the interest of, economy of syllable structure, the complex -sounds are analyzed as unitary elements. It is possible that appealing to, the Principle of economy of syllable structure may not help us in determining how to analyze complex or Suspicious segments. In such a situation, the analyst will,have to appeal to the principle of economy of 1B can be identifieg is CV. The forms in phonemes: lL. HYPOTHETICAL LANGUAGE 1. mtos |, ‘man’ 5., ftaft ‘uncle’ 2. smump . ‘dog’ 6. tfoms ‘house’ 3... tuf ‘child’ 7... stast,. ‘father’ | 4. mfom. | ‘woman’ 8. tsats . ‘mother’ it The forms in (1-4) show that one of the poet syllable structure in this set of data is CCVC, the forms js CCVC! ‘and 5 show that another clear-cut syllable structure is CC Scanned with CamScanner ay——_ Bitton 6 the example (3 ad finally, the ple (3) shows that the syllable js also-attested in the language. eve ei ay be analyzed as an See ue Complex segment ts in (8) AY cate, in Which case'the will have a syllable structure CVC which is obtainayye ome data. The suspicious segment may also be in the sequence of consonants, in which case the word will toa syllable structure CCVCC which is also obtainable-in the data. Therefore economy of syllable structure will be of no help i deciding one way or the other. At this point we will each appeal to the principle of economy of Phonemes.:If ts. is considered an affricate, it- will: introduce an extra phoneme which will not be economical: If on the other hand, the suspicious segments are viewed as a sequence, it will not introduce segments ‘since both -[t] and [s] are independently established as phonemes | (see Nos.) 1 and 3 «for instance). Therefore, in the ‘interest of economy of phonemes, | the suspicious segment must be viewed as a sequence. Sometimes, in considering the’ status of suspicious segments, it may be necessary to look at the: phonotactics of the language. Phonotactic rules are the principles which govem the arrangement of phonemes ‘in relation to another. For instance, in English, if such a chister must be the voiceless stop while the last consonant may be a liquid ({r] or [1]) as in words like split, stray, screech etc. The data below, from English will serve . illustrate the appeal to phonotactics: 12. ENGLISH kai ign cv ' ‘car? CV Bo ski: si 2) pep itgap’s evel Wo sto: ‘store’ ore 3) kat ~ ‘cat? cvc 15 skin ae ccve 4 sit ‘sit? cve 6 sti:l a cocy 5 dip sdipy eve ttt Scanned with CamScannerSE ON 70 _FRANCIS OYEBADE splei ‘splay’? =CCCV Roget 6 ‘at ve 7 strip ‘strip’ cccve 19 eekt ‘act’ vee 8 split ‘split?’ ~~ cccve, a Sxpks “thas” CVegg 8. Kemld ‘cold? cvcc dead judge ye P my ‘rouge’ cve |p eid? ‘rage’ ye p Sst ‘shin? = cve iit ae EY BR spepk ‘spank? Ccvce A tits “teach ye the complex Segments as a sequence of Consonants since each of the conso: nants [t], (4), [8] (16), [d] (5) and [2] are independently available in the language. However, the phonotactics of the language allows [t8] and [dz] to behave ; like unitary elements: a) they pattern like other Consonants in that they can occur word-initially, medially and finally. b) they may form minimal Pairs with other consonants (as it [tip] and [ti:t8]) Scanned with CamScanner“PHONOLOGY: 71 in the same 'se i i they always occur in ‘quential relationship: 7 sd’ always comes before [2] ‘and [t] before leit re sequence of consonants, it is reasonable to expect that o; = or the other may precede or follow. \ = Therefore even though it violates the principle of economy of phonemes, the phonotactics of English Suggest that the suspicious segments should be analyzed as affricates that is; unitary segments. jie Practice Exercise 250! 1. Identify the phonemes and allophones in the data below: 2.” State the distribution of the’ allophones a 3.\ Determine the status of complex segments: in the data: 1 [tindo}: ‘arm’ = T2-% [zaska] © ‘jar’ 2. [saska] ‘to weep’ 8. [totsi] “four” 3. [kiksi] ‘to smile’ 9:-[sakni] ‘to wipe’ 4. [nango] ‘water’ 10. [totko] ‘hound’ 5. [ninda] ‘soil’ 11. [zosta]’’ ‘not yet” 6. — [nossi] ‘human’ 12. [singa] ‘to want’ 1.6 Phonemic Analysis of Tones Some languages have “.....lexically significant, Contrastive but relative pitch on each syllable” (Pike 1957:3). These languages are called Tone Languages because “pitch Phonemes and segmental phonemes enter into the composition at least’ some morphemes” (Welmers 1973:80). Howevel; Just as it is done for consonant and vowel segments, it Is hn Recessary to determine the phonemic status of ; Odd Configurations in languages which use pitch contrastive'y- Scanned with CamScanner72, FRANCIS OYEBADE instance, the concept of minimal pairs can be used to estab] lish the phonemic status of tones in Yorbba. rin ‘to laugh’ min ‘to walk’ fin ‘to grate’ Any two of the words above form a minimal pair. Gen, only the tones on the lexical items which are obviously made up of the same segmental phonemes discriminate Meanings, Thus Yorbbé has three phonemic tones; High (represented as acute accent ["], Low (represented with grave accent CY) ana Mid (marked with an overbar (-) in phonological analysis but which by convention is unmarked in regular orthography), There are two other gliding tones which can be observed on the last syllable of the following illustrative utterances: kiré ‘move away’ pipd ‘plenty’ m6t6 ‘vehicle’ iyA ‘mother’ iwé ‘book’ Wi ‘town’ These gliding tones are the rising tone (represented by ["] ad the falling tone. ["]. These tones have predictable occurren® whereby the falling tone occurs only after a high tone while t | Tising tone occurs only after a low tone. Since these two tons | seem fo be distributionally predictable, they must be allophot | ‘ variants of some other tone(s). In order to determine what 100% | they are variants of, we need to examine more data: Scanned with CamScannerFE a nt tigo! L_] 55 “bottter 2. aga Lea) Sa eae 3, ilu LA 5.51 ‘town’ a Te FL) 3. ‘salt 5. 9mo --] 3.3. ‘child’ 6. ile ie Sl. 33 ‘house’ 7. kpukpo [ \] 1.15 ‘plenty’ 8. gbere [ -] 1.3 ‘indefinitely’ Of dudu (ee) a black [7] =1 = High {_]=5 = Low [-]=3 = Mid [/] = 51 = Rising Glide Tone [\] = 15 = Falling Glide Tone. It has been established earlier (using minimal pairs) that Yoribé had three phonemic tones: High, Mid, and Low. Logically, Yoruba should allow the following tonal patterns on disyllabic words: LL LM LH ML MM MH HM HL HH (which is observed in (1) of the data) (which is observed in (2) of the data) (which is not attested) (which is observed in (4) of the data) (which is observed in (5) of the data) (which is observed in (6) of the data) (which is observed in (8) of the data) (which is not attested) (which is observed in (9) of the data) i Scanned with CamScanner