| |||||
| Rules for linking phenomena |
| Handout: Structure of rules for linking phenomena |
Connected speech I: Assimilation & deletion
Language pedagogical & phonological analysis:
The probably biggest challenge in teaching connected speech is that “connected speech” is comprised of an immense amount of individual details and that there are no good materials out there that put all this into a coherent, teachable system. This problem was addressed over the years and decades with intonation, with recent works like Celce-Murcia et al. (2011) or Chun (2002) providing a coherent practical and theoretical system, respectively. While it still needs quite some creativity in curricular structuring, being able to rely on such practical presentations of the phonological system makes the whole endeavor significantly easier. The two units on connected speech phenomena - especially the one on linking - present an attempt at a practical realization of a coherent pedagogical presentation of the phonology of connected speech.
The first 2 resources of the first unit can easily take an entire 90 minute session. They are designed to draw learners’ attention to many of the individual aspects that comprise connected speech (esp. resource 1), with the idea of being able to identify them in truly authentic discourse and to be able to categorize them in analytical, though still simplified terms (resource 2). For this task I created a corpus of 25 video scenes between 3 and 46 seconds long that use certain connected speech features in a syntactically coherent discourse context. These scenes are from American TV shows (Scrubs, Boston Legal, Malcolm in the Middle) and use very authentic language. This system is superior to giving students an individual (pair of) words that have this process since language learning works best with authentic discourse. However, due to the enormous cognitive and perceptual challenges of these tasks students are given a list with transcriptions of relevant phrases after they listened to all files 2 times in order to catch missing instances. For the curricular structure the first connected speech unit has the global listening discrimination and deals with more salient features, while the second unit tackles more intangible aspects.
One of the most clearly noticeable aspects of connected speech is coalescent assimilation, in which non-sonorant continuants fuse with a following palatal glide to produce a voiced or voiceless affricate, depending on the phonation of the underlying continuant. In resource 3, this was introduced by linking this connected speech process to words in which it is lexicalized already, as in issue/this year. One of the most frequently occurring contexts for coalescent assimilation is when modal or auxiliary verbs terminating in an alveolar plosive are followed by “you”, which was utilized in R4 & 5 in guided and free practice. Due to the relative salience of coalescent assimilation, regressive and progressive assimilation should be easier to teach later on since this general concept of sounds influencing each other and changing accordingly is, then, already established.
The following part of this unit puts several individual connected speech processes together under the common theme of “deletion” (elision). R5 starts this with initial /h/ or /ð/ deletion in pronominal forms and have, which was combined with syncope (vowel deletion in unstressed syllables of certain words) and reduction of of to /ə/ if preceded by a consonant (/kʌplə/ - couple of) or to /əv/ if preceded by a vowel (/hɪstriəv/ - history of) (which is, in this case, only vowel reduction without deletion). This is communicatively put together in 2.1, which should be supplemented with some oral extension tasks. Note that most features in this unit so far were already established with the initial listening corpus, as well as, to a good extent, through the previous units on prosody, so that learners were prepared and, in fact, waiting for the rules to finally come, which has obvious effects on their possibility to learn them.
As a next step a lesson which is not in the materials was added that dealt with t-elision in nt clusters (pedagogically referred to as “en-tee deletion”), which emerges when t is preceded by n and followed by a vowel or syllabic l ([ɫ̩]). This was started with environments in which students probably did this already and which they know from the intonation unit, i.e. tags such as hasn’t he, isn’t he, aren’t I, won’t he. Then the inter prefix was added (international, interfere) and in connection to that, generally, int+V (printer, winter), ent+V (twenty, enter), ont+V and ant+V (with the last two producing very few words, e.g. Toronto or advantage). Nt-deletion may be perceived as somewhat colloquial, but is nonetheless employed very frequently by native speakers. In my experience this causes immense comprehension problems in learners, so that at least for speech perception purposes it is very important to teach. After some controlled practice with the number 20 (twenty) and some guided and free practice, the last topic is started which addresses consonant cluster reduction more generally. This is particularly relevant for learners of L1s which permit, phonotactically, more consonant clusters than English, so that learners would simply read English spelling letter by letter. R 5 and 6 practice this in various steps, first in tackling the impermissible or unusual /ktʃn̩/ (action), /ptʃər/ (rupture) and /ktʃər/ (picture) clusters within words, then in discussing how this kind of reduction emerges when endings are added (fact-s, exact-ly, band-s, kind-nes, rest-less), and finally across words if the following word does not begin with /h, j, w, r/ or a vowel. Some free practice should be added at the end.
Connected speech II: Linking
Didactic & phonological analysis:
In unit 8 the various linking processes are combined with flapping and regressive assimilation. The latter two aspects, though of course also occurring on their own within words (['bɛɾər], ['pæŋkeɪk]), also emerge frequently across word boundaries and can, thus, easily be perceived as a linking process and therefore be taught under this general header. In the entire pedagogical literature the topic of linking is the one aspect of English phonology that receives hardly any meaningful attention. In ELT student books like Baker/Goldstein (2008) or Gilbert (2009) it is hardly dealt with at all and in teacher guides on pronunciation teaching like Avery/Ehrlich (1992: 81ff) or Celce-Murcia et al. (2011: 163ff), which informed the rules presented in this unit, only a few pages deal with linking (and most other connected speech processes) in an extremely scarce manner. However, it has been shown in the various discussions on rhythm that linking is a significant aspect of English phonology and with its large amount of individual rules very important to teach. I have found over the years that many learners have a tendency to perceive clear breaks between words as “articulating well”, maybe reinforced by possible L1 systems of using glottal stops before stressed vowels (e.g. in a German phrase like am anderen Ende, every word would be preceded by a glottal stop, implicitly marking word boundaries).
I have also found that the extent to which English is linked leads to immense comprehension problems, which is of course directly connected to English rhythm, as outlined in that context. While the need to teach linking is quite clear, the large amount of individual rules and phenomena and the fact that they are almost completely devoid of any inherent communicative function (in order to build a meaningful context) makes this the pedagogically most challenging aspect of the English phonological system. In addressing this fact I have created a system of rules in a kind of diagram that shows how all the rules relate to each other and a PowerPoint presentation that contains all the individual rules to be supplemented after certain parts of this unit (all reproduced in a condensed *pdf version).
A further difficulty in teaching linking is that inductive teaching (discovery leaning – as opposed to deductive, explanation-based teaching) is rather difficult to use with linking. Normally it is essential for effective learning that students are allowed to “[discover] how language works rather than being given information about language construction ‘on a plate’” (Harmer 2007: 79)[1] since such language exploration leads to genuine understanding (Lewis 1996: 165). Learning and teaching with this kind of cognitive depth is not only significantly more fun and motivating, but also, of course, aids retention. This is also due to the fact that discovering language takes considerably longer than being given an explanation, and requires materials to engage in, all of which, again, allowing for cognitive depth. Linking processes, however, are difficult to ‘discover’ because they are rather intangible in fluent discourse and because the rules are so complex that a lot of guidance is needed. Therefore, in this unit a patchwork combination of techniques is used in which simple exploration techniques, either more student-centered or more teacher-guided are employed, while the rules on the PowerPoint slides are presented in-between after predictions were made. Also, the initial diagram is shown before each new aspect so that students know how things relate to each other. Using input from the video scenes corpus is further meant to connect to language students processed before.
Normal consonant+vowel linking (R1) is kept brief and is combined with flapping (R2 & R3), but could be extended very well with reading texts in a regular ESOL syllabus. The rules on the slides were presented between task 1 and 2. The phonological reasoning for these rules is an extension of the English syllable structure rule that onsets must be maximal. If the weight requirements for stress (a stressed syllable needs to have two X positions) condition a given consonant to be part of a rhyme, while the rule that onsets must be maximal conflicts with this necessity, this consonant can be ambisyllabic (e.g. pity, in which the t is the coda of the rhyme, forming the second X position, while it is at the same time onset of the weak syllable). However, this is only so strictly the case in citation form since “[i]n most accents of English [...] the maximization of onsets and the possibility of ambisyllabicity are generalised across word boundaries in connected speech” (Giegerich 1992: 280). A specific (American English) case of C+C linking, flapping, was introduced with the examples on the slides put on the board and compared by students, followed by the rule. R2 and 3 use the system also used above in that practice is initiated by adding inflections to selected words (used for guided practice in R3 task 2), before R2 task 2 starts with linking as flapping beyond word boundaries, continued in R3. Further free practice could be included with names as in R3 and a list of thematically somewhat related fields, as in R3 task 1. Resource 4, in introducing vowel+vowel linking (with the rules being introduced after task 1.2) builds a context through the use of idioms. Idioms and phrasal verbs (as in R2) are another feature that distinguishes very advanced from native English and are an aspect heavily emphasized in advanced and proficiency English course books (e.g. Jones 2002). Since they are also an essential component of oral language, it makes sense to study them together with pronunciation. The idioms in R4 are put together under the larger theme of business, with some, of course, being semantically field-independent. These idioms are personalized in task 3 and are further utilized in free practice in R10.
R5, finally, initiates the third “block” on the overview slide, consonant+consonant linking. This is where regressive assimilation was incorporated since this happens with similar consonants, just like “regular” linking happens between geminate consonants or stops. With R5 students, again, have to pronounce the words with each other and see if they notice any particular type of linking. With some stop+stop cases students noted that the last consonant of the first word is not “pronounced completely”, which is a good base to explain unreleased stops (further practiced in unit 9). In the second category words with approximants (fall leaves, far reaching) were more obvious than the others, so that these words could be used as an analogy. The rules on the slide were introduced throughout the activity. Task 2 offers some controlled practice, which some students enjoy, while others find this task too artificial.
In teaching practice regressive assimilation turns out to be more engaging than regular C+C linking. With R6 students are reminded of the term assimilation and that it refers to some kind of sound change phenomenon and are asked what something like “re-gressive” “as-simila-tion” could possibly be. With this in mind and with their state of language and phonological awareness at this stage, most pairs and groups tend to arrive at most rules relatively well. With “good girl” one student once noted it sounds like “Google”, which could be used quite nicely as a general analogy in this lesson (['guːgɫ̩] – ['gʊgˌgɝɫ]). Also, one’s shadow is very helpful since assimilation makes it extremely similar to one shadow, but the relative duration of the alveo-palatal fricative in one’s shadow makes them perceptually different, emphasizing the rule (though, of course, some co-articulation may well be perceptually present as well). While R7 and R8 task 1 provide some more controlled practice, R8 task 2 makes this more communicative and introduces or re-emphasizes the stress and vowel quality difference between the positive and negative form of the modal can (the positive being unstressed /kən/ and strongly linked to I and the following verb, the negative being /kæn/ and receiving contrastive sentence stress). Finally, R9 is a (language) game utilizing linking and other connected speech processes, while R10 is a longer creative free speech activity. It should be said that with more time, a variety of other activities could be included, such as analyzing written texts for possible linking, identifying linking in longer listening activities, guided practice in simulating certain types of conversations with respective practice and monitoring for linking, possibly initiated by oral input.
[1] This is not to say that explanations have no place in teaching at all (Fortune 1992: 168) and there are, of course, numerous teacher elicitation and illustration techniques (e.g. Harmer 2007: 203ff, Celce-Murcia et al. 2011: Ch. 9), but they should not replace student engagement.
Language pedagogical & phonological analysis:
The probably biggest challenge in teaching connected speech is that “connected speech” is comprised of an immense amount of individual details and that there are no good materials out there that put all this into a coherent, teachable system. This problem was addressed over the years and decades with intonation, with recent works like Celce-Murcia et al. (2011) or Chun (2002) providing a coherent practical and theoretical system, respectively. While it still needs quite some creativity in curricular structuring, being able to rely on such practical presentations of the phonological system makes the whole endeavor significantly easier. The two units on connected speech phenomena - especially the one on linking - present an attempt at a practical realization of a coherent pedagogical presentation of the phonology of connected speech.
The first 2 resources of the first unit can easily take an entire 90 minute session. They are designed to draw learners’ attention to many of the individual aspects that comprise connected speech (esp. resource 1), with the idea of being able to identify them in truly authentic discourse and to be able to categorize them in analytical, though still simplified terms (resource 2). For this task I created a corpus of 25 video scenes between 3 and 46 seconds long that use certain connected speech features in a syntactically coherent discourse context. These scenes are from American TV shows (Scrubs, Boston Legal, Malcolm in the Middle) and use very authentic language. This system is superior to giving students an individual (pair of) words that have this process since language learning works best with authentic discourse. However, due to the enormous cognitive and perceptual challenges of these tasks students are given a list with transcriptions of relevant phrases after they listened to all files 2 times in order to catch missing instances. For the curricular structure the first connected speech unit has the global listening discrimination and deals with more salient features, while the second unit tackles more intangible aspects.
One of the most clearly noticeable aspects of connected speech is coalescent assimilation, in which non-sonorant continuants fuse with a following palatal glide to produce a voiced or voiceless affricate, depending on the phonation of the underlying continuant. In resource 3, this was introduced by linking this connected speech process to words in which it is lexicalized already, as in issue/this year. One of the most frequently occurring contexts for coalescent assimilation is when modal or auxiliary verbs terminating in an alveolar plosive are followed by “you”, which was utilized in R4 & 5 in guided and free practice. Due to the relative salience of coalescent assimilation, regressive and progressive assimilation should be easier to teach later on since this general concept of sounds influencing each other and changing accordingly is, then, already established.
The following part of this unit puts several individual connected speech processes together under the common theme of “deletion” (elision). R5 starts this with initial /h/ or /ð/ deletion in pronominal forms and have, which was combined with syncope (vowel deletion in unstressed syllables of certain words) and reduction of of to /ə/ if preceded by a consonant (/kʌplə/ - couple of) or to /əv/ if preceded by a vowel (/hɪstriəv/ - history of) (which is, in this case, only vowel reduction without deletion). This is communicatively put together in 2.1, which should be supplemented with some oral extension tasks. Note that most features in this unit so far were already established with the initial listening corpus, as well as, to a good extent, through the previous units on prosody, so that learners were prepared and, in fact, waiting for the rules to finally come, which has obvious effects on their possibility to learn them.
As a next step a lesson which is not in the materials was added that dealt with t-elision in nt clusters (pedagogically referred to as “en-tee deletion”), which emerges when t is preceded by n and followed by a vowel or syllabic l ([ɫ̩]). This was started with environments in which students probably did this already and which they know from the intonation unit, i.e. tags such as hasn’t he, isn’t he, aren’t I, won’t he. Then the inter prefix was added (international, interfere) and in connection to that, generally, int+V (printer, winter), ent+V (twenty, enter), ont+V and ant+V (with the last two producing very few words, e.g. Toronto or advantage). Nt-deletion may be perceived as somewhat colloquial, but is nonetheless employed very frequently by native speakers. In my experience this causes immense comprehension problems in learners, so that at least for speech perception purposes it is very important to teach. After some controlled practice with the number 20 (twenty) and some guided and free practice, the last topic is started which addresses consonant cluster reduction more generally. This is particularly relevant for learners of L1s which permit, phonotactically, more consonant clusters than English, so that learners would simply read English spelling letter by letter. R 5 and 6 practice this in various steps, first in tackling the impermissible or unusual /ktʃn̩/ (action), /ptʃər/ (rupture) and /ktʃər/ (picture) clusters within words, then in discussing how this kind of reduction emerges when endings are added (fact-s, exact-ly, band-s, kind-nes, rest-less), and finally across words if the following word does not begin with /h, j, w, r/ or a vowel. Some free practice should be added at the end.
Connected speech II: Linking
Didactic & phonological analysis:
In unit 8 the various linking processes are combined with flapping and regressive assimilation. The latter two aspects, though of course also occurring on their own within words (['bɛɾər], ['pæŋkeɪk]), also emerge frequently across word boundaries and can, thus, easily be perceived as a linking process and therefore be taught under this general header. In the entire pedagogical literature the topic of linking is the one aspect of English phonology that receives hardly any meaningful attention. In ELT student books like Baker/Goldstein (2008) or Gilbert (2009) it is hardly dealt with at all and in teacher guides on pronunciation teaching like Avery/Ehrlich (1992: 81ff) or Celce-Murcia et al. (2011: 163ff), which informed the rules presented in this unit, only a few pages deal with linking (and most other connected speech processes) in an extremely scarce manner. However, it has been shown in the various discussions on rhythm that linking is a significant aspect of English phonology and with its large amount of individual rules very important to teach. I have found over the years that many learners have a tendency to perceive clear breaks between words as “articulating well”, maybe reinforced by possible L1 systems of using glottal stops before stressed vowels (e.g. in a German phrase like am anderen Ende, every word would be preceded by a glottal stop, implicitly marking word boundaries).
I have also found that the extent to which English is linked leads to immense comprehension problems, which is of course directly connected to English rhythm, as outlined in that context. While the need to teach linking is quite clear, the large amount of individual rules and phenomena and the fact that they are almost completely devoid of any inherent communicative function (in order to build a meaningful context) makes this the pedagogically most challenging aspect of the English phonological system. In addressing this fact I have created a system of rules in a kind of diagram that shows how all the rules relate to each other and a PowerPoint presentation that contains all the individual rules to be supplemented after certain parts of this unit (all reproduced in a condensed *pdf version).
A further difficulty in teaching linking is that inductive teaching (discovery leaning – as opposed to deductive, explanation-based teaching) is rather difficult to use with linking. Normally it is essential for effective learning that students are allowed to “[discover] how language works rather than being given information about language construction ‘on a plate’” (Harmer 2007: 79)[1] since such language exploration leads to genuine understanding (Lewis 1996: 165). Learning and teaching with this kind of cognitive depth is not only significantly more fun and motivating, but also, of course, aids retention. This is also due to the fact that discovering language takes considerably longer than being given an explanation, and requires materials to engage in, all of which, again, allowing for cognitive depth. Linking processes, however, are difficult to ‘discover’ because they are rather intangible in fluent discourse and because the rules are so complex that a lot of guidance is needed. Therefore, in this unit a patchwork combination of techniques is used in which simple exploration techniques, either more student-centered or more teacher-guided are employed, while the rules on the PowerPoint slides are presented in-between after predictions were made. Also, the initial diagram is shown before each new aspect so that students know how things relate to each other. Using input from the video scenes corpus is further meant to connect to language students processed before.
Normal consonant+vowel linking (R1) is kept brief and is combined with flapping (R2 & R3), but could be extended very well with reading texts in a regular ESOL syllabus. The rules on the slides were presented between task 1 and 2. The phonological reasoning for these rules is an extension of the English syllable structure rule that onsets must be maximal. If the weight requirements for stress (a stressed syllable needs to have two X positions) condition a given consonant to be part of a rhyme, while the rule that onsets must be maximal conflicts with this necessity, this consonant can be ambisyllabic (e.g. pity, in which the t is the coda of the rhyme, forming the second X position, while it is at the same time onset of the weak syllable). However, this is only so strictly the case in citation form since “[i]n most accents of English [...] the maximization of onsets and the possibility of ambisyllabicity are generalised across word boundaries in connected speech” (Giegerich 1992: 280). A specific (American English) case of C+C linking, flapping, was introduced with the examples on the slides put on the board and compared by students, followed by the rule. R2 and 3 use the system also used above in that practice is initiated by adding inflections to selected words (used for guided practice in R3 task 2), before R2 task 2 starts with linking as flapping beyond word boundaries, continued in R3. Further free practice could be included with names as in R3 and a list of thematically somewhat related fields, as in R3 task 1. Resource 4, in introducing vowel+vowel linking (with the rules being introduced after task 1.2) builds a context through the use of idioms. Idioms and phrasal verbs (as in R2) are another feature that distinguishes very advanced from native English and are an aspect heavily emphasized in advanced and proficiency English course books (e.g. Jones 2002). Since they are also an essential component of oral language, it makes sense to study them together with pronunciation. The idioms in R4 are put together under the larger theme of business, with some, of course, being semantically field-independent. These idioms are personalized in task 3 and are further utilized in free practice in R10.
R5, finally, initiates the third “block” on the overview slide, consonant+consonant linking. This is where regressive assimilation was incorporated since this happens with similar consonants, just like “regular” linking happens between geminate consonants or stops. With R5 students, again, have to pronounce the words with each other and see if they notice any particular type of linking. With some stop+stop cases students noted that the last consonant of the first word is not “pronounced completely”, which is a good base to explain unreleased stops (further practiced in unit 9). In the second category words with approximants (fall leaves, far reaching) were more obvious than the others, so that these words could be used as an analogy. The rules on the slide were introduced throughout the activity. Task 2 offers some controlled practice, which some students enjoy, while others find this task too artificial.
In teaching practice regressive assimilation turns out to be more engaging than regular C+C linking. With R6 students are reminded of the term assimilation and that it refers to some kind of sound change phenomenon and are asked what something like “re-gressive” “as-simila-tion” could possibly be. With this in mind and with their state of language and phonological awareness at this stage, most pairs and groups tend to arrive at most rules relatively well. With “good girl” one student once noted it sounds like “Google”, which could be used quite nicely as a general analogy in this lesson (['guːgɫ̩] – ['gʊgˌgɝɫ]). Also, one’s shadow is very helpful since assimilation makes it extremely similar to one shadow, but the relative duration of the alveo-palatal fricative in one’s shadow makes them perceptually different, emphasizing the rule (though, of course, some co-articulation may well be perceptually present as well). While R7 and R8 task 1 provide some more controlled practice, R8 task 2 makes this more communicative and introduces or re-emphasizes the stress and vowel quality difference between the positive and negative form of the modal can (the positive being unstressed /kən/ and strongly linked to I and the following verb, the negative being /kæn/ and receiving contrastive sentence stress). Finally, R9 is a (language) game utilizing linking and other connected speech processes, while R10 is a longer creative free speech activity. It should be said that with more time, a variety of other activities could be included, such as analyzing written texts for possible linking, identifying linking in longer listening activities, guided practice in simulating certain types of conversations with respective practice and monitoring for linking, possibly initiated by oral input.
[1] This is not to say that explanations have no place in teaching at all (Fortune 1992: 168) and there are, of course, numerous teacher elicitation and illustration techniques (e.g. Harmer 2007: 203ff, Celce-Murcia et al. 2011: Ch. 9), but they should not replace student engagement.
