Lenition and its Accompanying Constraints - Essay Example

Lenition and its Accompanying Constraints Introduction The German, Rudolf Thurneysen, is widely considered to have put the term lenition into first use in the late 19th century. Thurneysen noted that every case fundamentally involved a certain decrease in intensity of articulation, and because the state of Celtic studies required international terms, he had to suggest lenition, which was the Irish equivalent for aspiration (Thurneysen 1898: 43). The Celtic studies needed the se of international terms, and, as such, Thurneysen put forth two alternativbe words: lenierung and lenition. Lenition has played a significant role in the history of phonology. Although the term has been in existence since the 1950’s, it was not until the 1970’s that the gained common phonological discussion. Currently, the term has acquired a firm hold in the discourse of theoretical and historical phonology (Honeybone 2007).This paper seeks to analyze the various definitions of lenition as well as its accompanying constraints. The paper will also attempt to explain why lenition cannot occur in all positions in a word or phrase. Definitions of Lenition There are many different definitions of lenition. This is because quite a number of scholars have grappled with the concept of lenition. Phonologists even disagree on what types of segmental changes qualify as lenition. Nevertheless, various scholars have put forth their definitions of the term. Most definitions; however, imply that lenition and weakening are one and then same thing. Odden (2005: 239) avers that lenition is the interaction between consonants and vowels. Oden further states that it includes voicing of voiceless stops or voicing and sprantinization of stops. Lenition refers to the synchronic and diachronic sound alterations. In lenition, sounds may become weaker or weak sounds assume an allomorphic relation to stronger sounds. The weakening of sounds has been an area of concern for most scholars. Lenition has and continues to draw great literature over the decades. However, the general idea is that it implies some reduction in constriction degree or duration. Two other scholars, Ashby and Maidment (2005), seem to support Odden’s view of lenition. According to Ashby and Maidment, lenition is “also called weakening” (2005: 141). They go ahead to plot consonants on the scales of strength: a consonant is stronger the more it differs from a vowel, but becomes weaker when it begins to resemble a vowel. The idea of grading consonants on a scale of strength also features prominently in Trask’s Dictionary of Historical and Comparative Linguistics. Trask states: “A shift in character from left to right along any of the scales…may be regarded as lenition…” (Trask 2000: 190). The considerations of articulatory efforts may also lead to the concept of lenition. It is generally accepted that what causes the sound alterations that result into lenition is laziness. Scholars also observe that many of the world’s languages have undergone many changes. These changes find relevance in the context of weakening or lenition (Hock 1991: 80). However, this explanation has not yet found its way into a formal characterization of the patterns. Instead, sound patterns find expression in terms of language-specific rewrite canons. Then, there are those who treat lenition as a feature-spreading concept. This occurs under the framework of Autosegmental phonology where there is an obvious attempt to scale down lenition to a feature-spreading assimilation. According to Harris (1994), Spanish spirantization adheres to a rule that spreads from an adjacent segment. This concept is referred to as continuant. (Lombardi 1991). However, this approach faces various setbacks. To begin with, some forms of lenition, such as, degemination, debuccalization and elision only exist in autosegmental theory as deletion of phonological materials, and not as spreading. For instance, there is no feature that can be spread onto a consonant to turn it into ᴓ. This implies that the feature-spreading approach does not make room for a unified characterization of lenition. Lenition can also be considered as a scalar promotion. This is because lenition often involve chain shifts and a diachronic gradual weakening of stops into more reduced consonants. This often culminates into elision, creating the scalar nature of lenition. Two scales come to the fore: the sonority scale (Dell and Elmedlaoui 1985) and the strength scale (Hock 1991 and Lavoie 1996). In fact, for some scholars, any changes in the articulation of consonants amount to lenition. Such changes include then Spirantization of stops, debuccalization, opening of fricatives into approximants and outright deletion. An example of Spirantization of stops is /b/ > [β] while /β/ > [v] represents open fricatives mutating into approximants. /f/ > /h/ denotes debuccalization whereas deletion can be represented by /h/ > [0]. In spirantization of stops, debuccalization and opening of fricatives into approximants, there is a change from more open articulation to more closed one. The same is true for deletion if the deleted segment comprises of an oral articulation. The articulation may open more steps as is evident in Spanish in which voiced stops have the ability to lenite to frictionless approximants. Although lenition involves consonants, researchers have noted that vowel openness influences lenition. It appears that consonants are more likely to lenite when in the context of more open vowels (Kirchner 1998). If this were not true, it would be an indication that speakers seek to transport articulators for shorter distances in order to expend less effort. Moreover, open vowels cause consonants articulations to undershoot in one way or another; thus, enabling them to lenite more easily than consonants in the context of closed vowels. Observation arise from the definitions of by the afore-mentioned scholars. Odden establishes a relationship between lenition and an intersonorant context whereas Ashby and Maidment introduce the idea of consonant strength scales. Consequently, some consonants have more strength than others do. This paper notes that in all of these definitions of lenition, the bottom line is that lenition is the same as weakening of sounds. It is also important to note that lenition is a wide and complex concept involving many other ideas, such as spirantization, germination and degemination, buccalization and debuccalization, aspiration and affrication, among many others. As if that is not enough, many phonological constraints accompany lenition. Lenition and its Accompanying Constraints There are several processes of lenition. These are spirantization, degemination, debuccalization, voicing and flapping. Spirantization evolves from the word spirants. It occurs when stops lenite to form fricatives (Kenstowicz 1994: 35 -4398). The process of spirantization is common among Spanish dialects. This paper lays focus on Castilian Spanish. This dialect has both voiced and voiceless stops. There are voiceless stops in minimal pairs such as /bino/ and /pino/ (wine and pine) and /manga/ and /manka/ (sleeve and one-handed). Voiced stops, on the other hand, occur in in various positions, for example, utterance-initial and after nasals. This is especially the case with consonants b, d and y. In Spanish, voiced stops do not occur after continuant segments. This is evident is context-sensitive constraints [Ao] voiced stop and [Af] voiced stop. [Ao] refers to a voiced stop that is prohibited after an open segment. This implies that vowels and approximants are open (Steriade 1994) while [Af] is a voiced stop banned after a fricative. These two constraints are espoused in Steriade’s Aperture Theory (Steriade 1994). What motivates the two constraints is articulation because it is difficult to attain the closure needed for a voice stop without closing the surrounding segments (Kirchner 1998). It is worth noting that Spanish voice stops allow for utterance-initially and after segments. This implies that it is easier to prohibit a voiced stop after a fricative segment than preventing an approximant. Another case for lenition is flapping, a phenomenon otherwise known as tapping although flapping is more common than tapping (Honeybone 2007). Flapping refers to a synchronic process that has been innovated into modern English. It also covers Irish and most American dialects. Flapping only takes place in /t/ and /d/. Flapping in /t/ and /d/ causes the alveolar flap [ɾ], which is a lenis segment that can be indicated as: t, d → ɾ. Less constrained environments also experience flapping which cuts across word boundaries. Various examples abound; however, cases that are more restricted come to the fore. These include writing → [ˈɹaɪɾɪᶇ], madam → [ˈmaɾǝm], among others. The change from write [ˈɹaɪt] to writing [ˈɹaɪɾɪᶇ] is a clear indication that flapping is a synchronically active process. The concept of flapping has spread all over the world, effectively competing with other processes that affect the achievement of /t/. Some scholars (Harris 1994 and Kallen 2005) have characterized several processes operating in different dialects as causing the lenition of /t/. A good example is in the Liverpool dialect in which the lenition of /t/ clearly occurs. Apart from Spirantization and flapping, there are some cases in English that involve laryngeal change. This is the case for voicing. Such cases mark the transition from old to Middle English and exemplify changes that affect all fricatives in certain varieties of English. Examples of voicing include [f] and [v] and [s] and [z] as they occur in the following words: fader and vader (father) and synne and zenne (sin). A different type of laryngeal change also exists. A type of laryngeal change considered lenition in known as devoicing. Devoicing is evident in many languages including Catalan, Dutch, German, and Russian. This is in spite of the fact that it goes against the lenition hierarchy – it goes from voiced to voiceless. However, it is still lenition because it takes place in classic lenition environments: word-final. These examples can shed light on this phenomenon: actif and active, soules and soulez, and knowleche and knowledge. However, this is a controversial development in phonological theory. Nevertheless, it still makes sense as lenition because any process that alters laryngeal specification segments can be treated as voicing or devoicing. Salmons (2006) presents a dissenting analysis in the form of laryngeal realism. He maintains that obstruents of English, for example, do not make a characteristic of this feature. This is because fortis stops aspirate, and assimilation to lenis segments fails to take place. Lenition occurring closed syllables under laxed conditions experience three types of phonological constraints. These are *ClosedTense *[+tense, -low]C, *OpenLax *[-tense, -low]C and IDENT- [tense]. *ClosedTense occurs when tense vowels are restricted in closed syllables. It applies to non-low vowels, in which case it can result into the laxing of non-low vowels in closed syllables. The OpenLax constraint, on the other hand, occurs when non-low lax vowels get permission in open syllables: [-tense, -low]. Applying *OpenLax constraint causes the tensing of all non-low vowels in open syllables. Looking at these constraints, it is clear that they can cause an alteration in the tenseness of vowels from the input to the output. Consequently, there is need for the third constraint, IDENT-[tense]. This constraint ensures that tense values remain constant from input to output: αtense →αtense. Since IDENT-[tense] prevents possible changes in the tense value of the values, it receives prominence above the first two constraints. The tables below illustrate the domination of *ClosedTense and *OpenLax by IDENT-[tense]: /el/ el “the” candidates *ClosedTense *OpenLax IDENT-[tense] [el] *! →[ԑl] * /el/ el “the” candidates *ClosedTense *OpenLax IDENT-[tense] [el] *! →[ԑl] /le/ le dative pronoun candidates *ClosedTense *OpenLax IDENT-[tense] →[le] [lԑ] *! Another common constraint in lenition is the coda-cond. In order to appear in a coda, obstruents need to receive permission from the onset that follows. If a coda-cond constraint forces alterations to the input, it should rank above the IDENT constraint (McCarthy and Prince 1999). The implication of the outranking of IDENT by coda-cond is that some changes must be effected on the output to ensure no infringement on the latter. Since coda-cond permits the licencing of moraic obstruents by following onsets, it is possible to tackle the obstruents by getting rid of the segment and linking it the empty mora to the following onset in order to create a germinate. However, three other constraints would violate this rule. The following table shows the domination of coda-cond over IDENT: /mes.ke/ mes que “month that” candidates Coda-cond IDENT →[mԑk.ke] * [mԑs.ke] *! As stated above, other constraints violate the superiority of coda-cond to IDENT. One such constraint is MAX. This constraint states that it is not necessary to delete any input. The other constraint is Integrity, which states that one input can only result into one output. The third is *Gem. *Gem maintains that one segment may not be linked to both a coda and an onset. The violation occurs because, in coda-cond, as demonstrated by mes que (month that) above, there is the loss of /s/ and its subsequent replacement by /k/ in its output. This runs contrary to the MAX constraint, which holds that it is unnecessary to delete any input. As for the Integrity constraint, one input, which is /mes.ke/ should be equal to only one output. For the case of mes que, this is not possible since the input differs from the output. The input is /mes.ke/ while the output is /mek.ke/. It is possible to delete /s/ in coda-cond and replace it with /k/. However, the question is: is it possible to delete a mora? It may be easier to delete the entire coda that to delete a segment and then geminate the following onset. Deleting the coda would result into [me.ke], which is not a possible output of /mes.ke/. To solve this anomaly, another constraint has to come into play, the MAX-μ. The MAX-μ holds that any mora present in the input must also be available in the output. To ensure no violation of the MAX-μ, its ranking has to be higher than that of *Gem, Integrity and MAX-seg. To prevent outputs that are not possible from the given inputs, another constraint comes to the fore. This constraint is known as DEPons and it states that no onset absent from the input may appear in the output. For this constraint to work, it must be dominant over all the other constraints. In cases where there is the possibility of an undesired aspiration, another constraint is needed. Aspiration should be permitted only when there is no following onset. But when there is a following onset, there emerges need to prevent aspiration from occurring. The MAX-[place] constraint can prevent aspiration from taking shape. This constraint states that it may not be necessary to delete place features from the input. For aspiration to take place, place features have to give way, but MAX-[place] prevents this by ensuring that place values remain in place even in output. Using the example of mes que “month that”, it is important to ensure that /s/ is not rendered sonorous in the output. The /s/ should be silent in the output otherwise, the result would be impossible. For this reason, another constraint is used: IDENT-[son]. This constraint maintains that sonorous values must be maintained in the input and output. To achieve this feat, IDENT-[son] should be superior to all other constraints. Therefore, the input will be /mes.ke/ but the output becomes /mek.ke/. This is because /s/ is no longer sonorous. The discovery that deleting the moraic /s/ in Spanish is always accompanied by either germination or aspiration shows that the relationship between s-lenition and vowel laxing is clear. This is because vowel laxing is possible in closed syllables. If you remove the moraic /s/ without the resultant aspiration or germination, there would be a counter-bleeding relationship between laxing and lenition (Kiparsky 1973). This implies that if lenition was to come before laxing, then laxing would be impossible. Therefore, laxing has to take precedence over lenition. Conclusion Thurneysen brought the concept of lenition into being in the late 19th century. Since then, many researchers have grappled with the concept from different angles. Lenition has; therefore, become a wide and complex concept. It varies from language to language and its definition varies from author to author. Consequently, it is a term that has generated many definitions. However, these definitions have a convergent point: lenition involves the weakening of consonant sounds in words. Lenition also has several accompanying constraints, which play an important role in its process. The constraints act as rules upon which lenition can take place. For this reason, lenition cannot occur in all places of a word or a phrase. It can either be at the beginning, word-initial, in the middle, medial, or at the end, word-final. In essence, to fully grasp the concept of lenition, it is important to study it from a broad perspective that brings in as many researchers as possible. Bibliography Ashby, M. and J. Maidment. (2005) Introducing Phonetic Science. Cambridge: Cambridge University Press. Dell, F. & M. Elmedlaoui (1985) Syllabic Consonants and Syllabification in Imdlawn Tashlhiyt Berber, J. African Languages and Ling. 7, 105-130. Harris, J. (1994) English Sound Structure. Oxford: Blackwell. Harris, J. (2003). Grammar-internal and grammar-external assimilation. University College London Working Papers in Linguistics, 15. Hock, H. H. (1991) Principles of Historical Linguistics. Second edition. Berlin: Mouton de Gruyter. Honeybone, P. (2007). ‘New-dialect formation in nineteenth century Liverpool: a brief history of Scouse’, in Anthony Grant and Clive Grey (eds.). The Mersey Sound: Liverpool’s Language, People and Places. Liverpool: Open House Press, 106-140. Honeybone, P. (2008). Lenition, weakening, and consonantal strength: tracing concepts through the history of phonology. In J. de Carvalho, T. Scheer, & P. Ségéral (eds.), Lenition and Fortition. Berlin: Mouton de Gruyter. 9-93. Kallen, J. (2005). ‘Internal and external factors in phonological convergence: the case of English /t/ lenition’, in P. Auer, F. Hinskens, and P. Kerswill (eds.), Dialect Change: the Convergence and Divergence of Dialects in Contemporary Societies. Cambridge: Cambridge University Press. Kiparsky, P. (1973) Abstractness, opacity and global rules. Three Dimensions of Linguistic Theory 57-86. Tokyo: TEC. Kenstowicz, M. (1994) Phonology in Generative Grammar. Cambridge, MA: Blackwell. Kirchner, R. (1998). An effort-based account of consonant lenition. Doctoral dissertation, University of California at Los Angeles. Kirchner, R. (2004). Consonant lenition. In B. Hayes, R. Kirchner & D. Steriade (Eds.), Phonetically based phonology (pp. 313-345). Cambridge, UK: Cambridge University Press. Lavoie, Lisa M. (2001). Consonantal strength: Phonological patterns and phonetic manifestations. New York: Routledge. Lombardi, L. (1999) Positional Faithfulness and Voicing Assimilation in Optimality Theory. Natural Language and Linguistic Theory 17: 267-302. McCarthy, John J. and Alan P. (1995). Faithfulness and Reduplicative Identity. University of Massachuetts Occasional Papers in Linguistics. 18 Papers in Optimality Theory: 249-384. Odden, D. (2005) Introducing Phonology. Cambridge: Cambridge University Press. Oxford English Dictionary, Oxford: Oxford University Press. Steriade, D. (1999). Phonetics in phonology: The case of laryngeal neutralization. In M. Gordon (Ed.), Papers in Phonology, 3 (pp. 25-146). Los Angeles: UCLA. Thurneysen, R. (1898) Review of Pedersen (1897). Anzeiger für indogermanische Sprachund Altertumskunde. Beiblatt zu den Indogermanischen Forschungen 9: 42-48. Trask, R.L. (2000) The Dictionary of Historical and Comparative Linguistics. Edinburgh: Edinburgh University Press. Read More