3.3 Pilot-one
3.3.1 Development of the instrument (toddler CDI in isiXhosa)
52 of words was counted to inform the word list for the first pilot. Since the transcriber (a fluent speaker of both English and isiXhosa) transcribed the children’s spontaneous speech utterances from a video file and not a sound file, she was able to use the linguistic and non-linguistic contexts of the conversations to determine the children’s target utterances. This was also aided by a set of notes I made during recording and by asking caregivers to repeat what their child was trying to say.
Transcriptions were further used to inform a table of grammar acquisition ‘errors’ (see Appendix F) and caregiver responses from the focus groups and interviews were documented alongside the translated version of the American English CDI to inform which words were commonly known.
53 household items’, ‘Furniture and rooms’, ‘Outside things’, ‘Places to go’, ‘People’, ‘Routines and games’, ‘Action words’, ‘Descriptive words’, ‘Words about time’, ‘Pronouns’, ‘Questions words’,
‘Prepositions and locations’, ‘Words about amounts’, and ‘Conjunctions/sentence modifiers’.
In completing the CDI, caregivers are required to say whether their child understands and says a word, and it was explained that it does not count if the child simply mimics a word used by an adult. Receptive abilities as a category of their own are not considered, as recommended by Dr Alcock, to shorten the tool as well as the fact that, according to Eriksson et al. (2012: 228), the comprehension component of the CDI (when it has been included as a category on its own) is a less precise measure than production.
For purposes of the qualitative component, when words are listed as synonyms (for example, the English word ‘want’ can be realised as funa, rhalela or bawela in isiXhosa) then caregivers are asked to indicate which of the synonyms their child knows. For example, for the word ‘baby’, the list would look like this if a caregiver gives ubhabha as the synonym the toddler uses:
If their child knows a different word with the same meaning, it is written next to the word on the CDI.
For example:
Caregivers are additionally asked to provide the child’s pronunciation of the word if they know it but are not yet able to give the standard pronunciation. For example:
If the child does not know the word ‘friend’ umhlobo/itshomi but has a friend whose name they know and call, they are awarded for that too. For example:
54 3.3.1.2 Sentences and Grammar
The table of grammatical constructions from children’s spontaneous speech (see Appendix F), caregiver responses, as well as existing literature on child acquisition were used to inform the grammar section of the CDI. The grammar section for all Bantu languages is created identical (except for two additional questions on the isiXhosa CDI),42 and it is thus also, to a certain extent, informed by Sesotho and Setswana pre-pilot data (the Xitsonga team joined the adaptation later and thus had not yet completed the pre-pilot stage).
This section is also informed by Dr Alcock’s findings from the grammar section of the Kenyan CDI. For example, she finds there is no correlation between a child’s age, nor with any measures from spontaneous speech recordings for verb forms such as the passive,43 so she recommended this be left off the Southern Bantu language CDIs. Dr Alcock additionally recommended that questions should not be asked about grammatical items that are not acquired in an errorless manner (for which the spontaneous speech was consulted), as this is less helpful in assessing development. This approach is similar to that of Connelly (1984: 66) who, in his study, depends on the analysis of children’s errors since, as mentioned earlier, “it is only through these that [children’s] progress towards adult forms can be plotted”. Dr Alcock also recommended that the CDI does not ask about verbal extensions since, for example, the applicative44 in Kiswahili and Kigiriama is either not used or used correctly. According to Suzman (1991), verbal extensions appear to be initially lexically learned and as Demuth (1992) says of the causative, treated as a lexical whole even by adults – although this may be less true for the applicative.
42 The first asks about the Cl. 9 NPx (which is zero in Sesotho and Setswana) and the second asks about the Cls 1 and 1a identificative copulative (which in Sesotho and Setswana is identical across all nouns classes).
43 See Suzman (1991) for an appraisal of the acquisition of the passive in isiZulu.
44 See Demuth (1998) regarding Sesotho-speaking children’s use of the applicative in the 2 to 3-year-old age group. Morphology of the applicative acquisition in Sesotho may be a more productive field of study due to the various forms in which it is realised in the language (very few allomorphs of the applicative occur in isiXhosa).
This being said, more research is needed on the acquisition of the applicative in isiXhosa regarding children’s ability to use it in the correct contexts (i.e. locative, benefactive), although this is beyond the scope of the grammar section of the CDI.
55 The grammar section is split into the following categories, of which I will discuss the construction of the questions relevant to the analysis in this thesis. See Appendix G for the full grammar section of the CDI.
Section A consists of the following yes/no questions:
• The production of noun class prefixes
For this question Cls 1 and 1a are used as examples, as recommended by Dr Alcock.
• The production of plurals
For this question Cls 1/2, 5/6, and 9/10 are used as examples, as recommended by Dr Alcock and as informed by the pre-pilot spontaneous speech examples.
• The production of correct adjective agreement
Adjectives 'big' and 'small' are used for examples with agreement markers of Cls 1, 9, and 5, as recommended by Dr Alcock and as informed by the pre-pilot spontaneous speech examples.
• The production of the present tense (disjunctive)
• The production of the recent past tense (disjunctive)
First and second person subject markers are used in the examples as spontaneous speech samples showed these were commonly known and used by this age group.
Section B asks respondents to pick the option that sounds most like their child’s speech at the time, in testing for the following grammatical components. Two or three options are given for each item, with the first being the most basic form, often with missing agreement and tense markers and the last being the full, appropriate adult form. The questions45 elicit information on:
45 All of these questions are either informed by the Kenyan CDIs, spontaneous speech samples, or both. Questions on the copulative below are only informed by observations of errors in spontaneous speech, as questions on copulative forms do not appear on the Kenyan CDIs.
56
• NPx production for specific classes
• The conjunctive past tense
• Subject marker omission in negative
• The production of the full negative imperative
• The correct use of the subjunctive
• The full or partial use of the future tense
• Present tense
• Production of object markers without tense marking
• Production of object markers with present tense
• Production of the copulative
• Production of the full or partial Cl. 1 copulative
• Production of the standard Cl. 5 copulative or an overgeneralisation of Cl. 9 copulative
Section C ask for the child’s three longest utterances, if they are combining words. If they are not yet combining words, only Sections A and B are completed.
Section D considers the complexity of the child’s multi-word utterances. This section is only completed if children are already combining words. For the purposes of the results in this thesis, I consider the following from this section:
• Complexity in the use of the possessive pronoun
• Complexity in the use of prepositions
Vocabulary items used in Sections B and D are informed by the pre-pilot research.
Possessives and possessive markers are asked under a category in the words section since pronouns like ‘mine’ are not considered part of grammatical acquisition in English, however, for the purposes of my results and analysis I consider them as grammatical due to the interesting results that the Bantu language agreement provides. I also analyse prepositions under the grammar section since, whilst words like ‘up’ and ‘outside’ are stand-alone words in English and are asked as such on the CDI, they are lexicalised morphologically in isiXhosa.
57 3.3.1.3 Family history questionnaire
During the workshops an existing family history questionnaire used by linguists at the University of Stellenbosch was adapted for the Southern African CDI, through consultation with key role players.
This questionnaire accompanied the CDI. The questionnaire and CDI instructions (see Appendices G and H) were translated into isiXhosa by one first-language isiXhosa speaker together with a second- language professional in the field. This questionnaire was pre-piloted before pilot-one fieldwork, with two caregivers from different SES in Cape Town (aged 60 and 23), to ensure the questions were eliciting the responses required and to avoid children’s scores being negatively impacted due to poorly translated questions.