The current study explores inference-making processes in 10–12-year-old bilingual Malawian children who either listened to stories in their primary language, or L1 (Chichewa), as compared to their secondary language (L2) (English), or viewed cartoon films containing no verbal content. The 127 children who participated in the study were divided into six groups characterised by different conditions of stimulus presentation – stimuli varied with respect to their modality (non-verbal film versus pre-recorded stories) or the language of stimulus presentation (English or Chichewa). The results indicate that the pre-recorded audio recordings seem to have supported inference-making more than the corresponding wordless films. This finding illustrates the significance of linguistically presented content. The linguistically presented content elicited even more inferences when it was presented in the children’s L1 (Chichewa) rather than in their medium of academic instruction (English). However, the results also indicate that the children from the private school (with English as a medium of instruction) drew more inferences than the children from the public school (where Chichewa is the medium of instruction). Furthermore, the results reveal that while the children were able to use knowledge transfer from a variety of knowledge bases to draw inferences, the inferencing process was impeded when the story content deviated too much from their own experiences. Lastly we found indications of variations in inferencing patterns that seemed to correlate to the language in which the stimulus material was presented and responded to.
In order to understand text, it is essential that comprehenders draw inferences, because generally a text can only provide a certain amount of essential information, which the recipients are required to expand with inferences. Inferences that serve to fill gaps identified in a text are generated through the use of different types of knowledge bases (Keenan, Baillet & Brown
There is an enormous number of inferences that comprehenders make; below we provide an overview of different types of inferences with the aim of positioning the subject of our current investigation within the research field.
There is hardly any consensus among researchers in psycholinguistics and discourse processing on the different types of inferences that readers, listeners or viewers make (Graesser, Singer & Trabasso
Distinctions between different types of inferences.
Authors | Identifying distinctions between types of inferences | |
---|---|---|
McKoon and Ratcliff |
Automatic | Strategic |
Graesser |
On-line | Off-line |
Graesser |
Text-connecting | Knowledge-based |
Graesser |
Local | Global |
Barnes |
Coherence | Elaborative |
Cain and Oakhill |
Text-connecting | Gap-filling |
Bowyer-Crane and Snowling |
Coherence | Elaborative/knowledge-based |
Cromley and Azevedo |
Anaphoric text-to-text | Background-to-text |
Graesser
Text-connecting inferences are needed to integrate single sentences into one cohesive textbase. Thus, they bridge gaps between different phrases and sentences on a linguistic level. A good example for a text-connecting inference is provided by Cromley and Azevedo (
Knowledge-based inferences embellish story content and amplify a story’s context. They provide a fuller representation of an event, but they are not central to textual cohesion (Barnes
In terms of the distinction between
Other categories of inferences that Graesser
In sum, we can now categorise knowledge-based or elaborative inferences as global, off-line, background-to-text inferences that operate over the general knowledge base of the comprehender and hence use long-term memory representations.
In addition to the debate on different categories of inferences, researchers also entertain either a ‘minimalist’ (e.g. McKoon & Ratcliff
The minimalist view holds ‘that there is only a minimal automatic processing of inferences’ during comprehension. McKoon and Ratcliff (
In contrast, the constructionist theory maintains that a comprehender may be engaged in a constant ‘search after meaning’ in order to build a situation model of a text or a visual stimulus that is coherent both at a local and at a global level. In this way, the comprehender will draw all the inferences needed to explain why things are mentioned in the text – or appear in a film – in order to achieve local and global coherence. Kintsch (
The CIM includes different levels of mental representation to integrate the propositions and/or ideas that make up the elements of a text representation. The CIM assumes that comprehenders construct a coherent mental representation of the text they are reading or listening to, that is, the
The situation model is a type of mental model according to, for example, Johnson-Laird (
The significance of the situation model lies with the fact that information is retained in memory for a certain period of time. Situation models are considered as providing ‘deeper’ levels of both representation and comprehension than other text-related representations (Graesser, Olde & Klettke
According to Rai
Rai
The construction of a coherent situation model may hence be hindered not only by the processing of current knowledge but also by a lack of pre-existing knowledge. Such knowledge is often stored in the form of schemas. In the section that follows, we explain the role that schemas play in comprehension.
Research in psychology and cognitive science demonstrates that our past experiences and our background knowledge play a crucial role in comprehension, some of which may be stored in the form of schemas
Kintsch and Greene’s (
Similar results, which seem to confirm that story schemas are culture-specific, generally demonstrate a positive relationship between cultural familiarity and reading or listening comprehension (Alptekin
The current study investigates knowledge-based (i.e. elaborative inferences) in 10–12-year-old bilingual Malawian children. Inferencing was investigated not only when the children listened to stories in their primary language, or L1 (Chichewa), as compared to their secondary language, or L2 (English), but also when they viewed wordless cartoon films. The study adopted a question-answering methodology to reveal whether the children used knowledge either from the previous text or from their knowledge base to amplify their situation model with inferred information. In particular, the participants answered
Our participants were learners at two different schools: a private school, Kapita Primary School, and a public school, Mphungu Primary School, which are both located in the Central Region of Malawi in the capital city, Lilongwe. The medium of instruction at Kapita Primary School is English in all grades except when studying a language such as Chichewa as a school subject. In contrast, Chichewa is the medium of instruction from grade 1 to grade 4 at Mphungu Primary School. The language policy in Malawi permits the use of a mother tongue from grades 1 to 4 in public schools, and English becomes the medium of instruction beginning in grade 5.
The total number of participants was 127 children (64 female, 63 male) aged between 10 and 12 years (44 10-year-olds, 40 11-year-olds and 43 12-year-olds). At the time of the study, 69 of the children were attending the private school (Kapita), while 58 of them were attending the public school (Mphungu).
Most of the participants were from the same urban community, while some of the participants who attended the public school were from rural areas. On average, the participants who attended the private school were from a higher socio-economic background than those who attended the public school.
Parents who had given consent for their children to participate in the study were provided with a language background questionnaire, in which we asked about the language repertoire of the children, the children’s involvement in storytelling, languages used in storytelling, the children’s knowledge of oral traditional practices and languages used when engaging in these oral traditional practices (see
Background information about children attending Kapita and Mphungu primary schools.
Background information | Kapita |
Mphungu |
||||
---|---|---|---|---|---|---|
Chichewa | English | Both English and Chichewa | Chichewa | English | Both English and Chichewa | |
Language(s) used in storytelling at home | 15 | 9 | 38 | 36 | 2 | 18 |
Ability to [re]tell stories in languages | 17 | 6 | 43 | 39 | 3 | 15 |
The elicitation material used in this study comprised two wordless cartoon films [‘The mole and the rocket’ (1966) and ‘The mole and the radio’ (1968), developed for children by the Czech cartoonist Zdĕnek Miler in 1956
An English verbal version of each of the cartoon films was constructed by a postgraduate student in Creative Writing. In terms of comparability to the film, the student adopted the measures used by Meringoff (
In a final step, the participants’ teachers assessed the stories in order to ensure that they were at an appropriate level of grammatical and lexical complexity for the participants.
The participants were divided into six groups. Each group of 20–21 children was assigned a different story condition (condition 1 (a) to condition 5). Conditions varied according to the modality of the stimulus material (film vs. aural stories) and the language of stimulus presentation (English or Chichewa). In all conditions, the first author and the local assistant tested the participants individually in a vacant room at their school. The children were briefed on their rights and signed the informed consent form. They were requested to pay close attention to the stimulus material and were told that we would ask comprehension questions based on the film or aural story.
Each child watched the cartoon film ‘The mole and the rocket’ (1966) and subsequently answered comprehension questions in English. After a 1-week time lapse, the participants watched the second cartoon film, ‘The mole and the radio’ (1968). Again they answered comprehension questions in English.
Each child watched the cartoon film (‘The mole and the rocket’) and afterwards answered comprehension questions in Chichewa. The children watched the next cartoon film, ‘The mole and the radio’ (1968), after a 1-week time lapse and subsequently answered comprehension questions in Chichewa.
Each child listened to a pre-recorded reading of the story ‘The mole and the rocket’ in English. After listening to the story, they answered comprehension questions in English. After a 1-week time lapse, the children listened to a pre-recorded reading of the second story, ‘The mole and the radio’, in English. After listening to the story, they answered comprehension questions in English.
Each child listened to a pre-recorded reading of the story ‘The mole and the rocket’ in Chichewa. After listening to the story, they answered comprehension questions in Chichewa. After a 1-week time lapse, the children listened to a pre-recorded reading of the second story, ‘The mole and the radio’ in Chichewa. After listening to the story, they answered comprehension questions in Chichewa.
Each child listened to a pre-recorded reading of the story ‘The mole and the rocket’ in Chichewa. However, after listening to the story, children answered comprehension questions in English. After a 1-week time lapse, the children listened to a pre-recorded reading of the second story, ‘The mole and the radio’, in Chichewa. Again they answered comprehension questions in English.
Each child listened to a pre-recorded reading of the story ‘The mole and the rocket’ in English. However, after listening to the story, the children answered comprehension questions in Chichewa. After a 1-week time lapse, the children listened to a pre-recorded reading of the second story, ‘The mole and the radio’, in English. Again they answered comprehension questions in Chichewa.
In particular, we asked the children the following six questions to elicit elaborative inferences.
Rocket story questions:
Why did the engine of the rocket stop? Why did the little mole not dive into the sea to look for the parts of the rocket himself? Why did the mole make friends with this sea creature?
Radio story questions:
What was the bee doing inside the flowers at the beginning of the story, and what was the bee doing inside the flowers towards the end of the story? Why did the radio make strange noises and then completely stop playing? Why did the mole wrap the scarf around the radio?
Each of the participants’ answers to the questions were transcribed and thereafter analysed. The following procedure was used in the analysis: All the answers that were based on information that was not explicitly stated in the stories were considered inferences. Such answers were kept for further analysis. In contrast, all the answers that were classified as text-based were put aside, as our study only focused on inferences. Thereafter, the inferences were further analysed in order to find the total number of inferences that were drawn in English and Chichewa by both groups of participants from each of the two stories. These inferences were arranged in accordance with the stated conditions that participants were assigned to. Further analysis was conducted in order to determine the total number of inferences drawn by each group of participants (Kapita vs. Mphungu) from each of the two stories. Further analysis enabled us to put together inferences that depicted the participants’ pre-existing knowledge bases and those that depicted the children’s use of knowledge transfer from other subjects.
Inferences in the participants’ answers were initially identified by the first author, and they were subsequently verified by the second author. The answers that were in Chichewa were verified by the first author’s colleague, a specialist in African Linguistics. For all the inferences that were identified, inter-coder agreement was high (97% and 98%, respectively). Cases of disagreements between coders were discussed and resolved.
Ethical clearance for the current study was granted by the University of KwaZulu-Natal’s Ethics Review Committee (no. HSS/0246/012D). We obtained formal written consent from the school principals, teachers, parents and children prior to the study.
In order to present our findings, we first compare the total number of inferences that participants drew when they were answering comprehension questions in their L1 with those they drew when they used their L2. We compare these results with information that shows that participants from Kapita Primary School generated more inferences than participants from Mphungu Primary School. Subsequently we contrast the total number of inferences drawn when the mode of presentation was aural (pre-recorded story) with those drawn when the mode of presentation was visual (film). Thereafter, we compare the total number of inferences from ‘The mole and the rocket’ (hereafter ‘the rocket story’) with those drawn from ‘The mole and the radio’ (hereafter ‘the radio story’). We also present results reflecting that participants were using knowledge transfer from a variety of school subjects when answering the comprehension questions.
The participants drew a total of 81 inferences from the radio story and 49 inferences from the rocket story when answering comprehension questions in English. However, they managed to draw a total of 97 inferences from the radio story and 78 inferences from the rocket story when answering comprehension questions in Chichewa (see
Total number of inferences drawn in English and Chichewa by both groups of participants from the radio story.
Question number | Story title | Condition (English) | Number of inferences (English) | Condition (Chichewa) | Number of inferences (Chichewa) |
---|---|---|---|---|---|
Question 1 | ‘The mole and the radio’ | 1 (a) | 11 | 1 (b) | 10 |
‘The mole and the radio’ | 2 | 17 | 3 | 14 | |
‘The mole and the radio’ | 4 | 14 | 5 | 19 | |
Question 2 | ‘The mole and the radio’ | 1 (a) | 9 | 1 (b) | 6 |
‘The mole and the radio’ | 2 | 5 | 3 | 7 | |
‘The mole and the radio’ | 4 | 6 | 5 | 6 | |
Question 3 | ‘The mole and the radio’ | 1 (a) | 10 | 1 (b) | 5 |
‘The mole and the radio’ | 2 | 6 | 3 | 11 | |
‘The mole and the radio’ | 4 | 3 | 5 | 19 | |
Total number of inferences drawn in English and Chichewa by both groups of participants from the rocket story.
Question number | Story title | Condition (English) | Number of inferences (English) | Condition (Chichewa) | Number of inferences (Chichewa) |
---|---|---|---|---|---|
Question 1 | ‘The mole and the rocket’ | 1 (a) | 11 | 1 (b) | 9 |
‘The mole and the rocket’ | 2 | 12 | 3 | 12 | |
‘The mole and the rocket’ | 4 | 10 | 5 | 11 | |
Question 2 | ‘The mole and the rocket’ | 1 (a) | 6 | 1 (b) | 8 |
‘The mole and the rocket’ | 2 | 7 | 3 | 6 | |
‘The mole and the rocket’ | 4 | 0 | 5 | 18 | |
Question 3 | ‘The mole and the rocket’ | 1 (a) | 2 | 1 (b) | 1 |
‘The mole and the rocket’ | 2 | 0 | 3 | 3 | |
‘The mole and the rocket’ | 4 | 1 | 5 | 10 | |
There were a number of inferences that were common regardless of the language used, for example, the engine of the rocket stopped working because it did not have enough oil or it stopped working because the mole was failing to control it. However, there were also inferences that seemed to be language-specific, for example when responding to the question, ‘what is the bee doing in the flower?’, some of the children included answers such as ‘the bee is pollinating the flowers’, ‘the bee is collecting nectar’ or ‘the bee is eating the nutrients found in flowers’ when the language of verbalisation was English. The children did not include such answers when answering the same question in Chichewa.
The data in
The results in this section suggest an interaction between the language of instruction and the capability to react to inference elicitation questions in this language. While both groups of children were L1 Chichewa speakers who had a significant amount of exposure to Chichewa at home according to the parental questionnaire (see
Conversely, when comprehension questions were answered in English the children who attended Kapita Primary School (i.e. private school) drew more inferences than those who attended Mphungu Primary School (i.e. public school) (see
A comparison of the total number of inferences by each group.
Condition | Results from the rocket story when using English (L2) |
Results from the rocket story when using Chichewa (L1) |
Results from the radio story when using Chichewa (L1) |
Results from the radio story when using English (L2) |
||||
---|---|---|---|---|---|---|---|---|
Kapita | Mphungu | Kapita | Mphungu | Kapita | Mphungu | Kapita | Mphungu | |
Condition 1 (a) | 7 | 4 | 4 | 5 | 3 | 7 | 7 | 4 |
Condition 2 | 9 | 3 | 8 | 4 | 9 | 5 | 12 | 5 |
Condition 4 | 7 | 3 | 5 | 6 | 10 | 9 | 8 | 6 |
L1, first language; L2, second language
The quantitative results presented in
Where answers were required in Chichewa, those children who attended the public Mphungu Primary School provided much greater detail in their answers than their peers from the private Kapita Primary School; compare example sentences 1 and 2, which were produced by participants attending the public Mphungu Primary School:
In comparison the children who attended the private Kapita Primary School struggled with verbalising their inferences in great detail in Chichewa. Their Chichewa answers typically consisted of short phrases and lacked specific names for referents as exemplified in examples 3 and 4:
The opposite holds for inferences that were verbalised in English. Here, the children who attended the private Kapita Primary School provided more detailed verbalisations of inferences than their peers from the public Mphungu Primary School. The level of detail of the English verbalisations from children attending the private Kapita Primary School is illustrated in examples 5 and 6:
‘It stopped because, the engine of the rocket stopped because there was no [sic] any oil or things to put into the engine.’ (Participant from Kapita Primary School, Female, Age 10)
‘It crashed because the mole didn’t know how to control it and where to go.’ (Participant from Kapita Primary School, Female, Age 12)
In comparison the children who attended the public Mphungu Primary School typically provided short answers that lacked specific names for referents; this is represented in examples 7 and 8:
‘I think it needed fuel.’ (Participant from Mphungu Primary School, Female, Age 12)
‘It was overspeeding.’ (Participant from Mphungu Primary School, Female, Age 11)
Our results thus indicate that the dominant language for the two sets of children may be determined by the type of school they are attending: Chichewa seems to have been the dominant language for the children attending the public Mphungu Primary School, whereas English seems to have been the dominant language for the children who were attending the private Kapita Primary School. The language background questionnaire supports this finding, as the children who attended Kapita Primary School seemed to have much less exposure to Chichewa at home than their peers from Mphungu Primary School (see
The results in this study show that participants drew more inferences when the mode of presentation was aural rather than visual (film) (see
Total number of inferences drawn from the two modes of presentation (aural and visual).
Question number | Story title | Condition (English) | Number of inferences (English) | Condition (Chichewa) | Number of inferences (Chichewa) | Total |
---|---|---|---|---|---|---|
1, 2, 3 | ‘The mole and the radio’ | 1 (a) | 30 | 1 (b) | 21 | 51 |
1, 2, 3 | ‘The mole and the rocket’ | 1 (a) | 19 | 1 (b) | 18 | 37 |
1, 2, 3 | ‘The mole and the radio’ | 2 | 28 | 3 | 32 | 60 |
1, 2, 3 | ‘The mole and the rocket’ | 2 | 19 | 3 | 21 | 40 |
1, 2, 3 | ‘The mole and the radio’ | 4 | 23 | 5 | 44 | 67 |
1, 2, 3 | ‘The mole and the rocket’ | 4 | 11 | 5 | 39 | 50 |
The total number of inferences for conditions 1 (a) and 1 (b) – where the stimulus was the non-verbal film – were 49 and 39, respectively, which was generally lower than the total number of inferences obtained when the mode of presentation was aural (compare the numbers in
The general pattern revealed in this table is that the number of inferences drawn from the visual mode (i.e. 88) was lower than those drawn from the aural mode of presentation (i.e. 100 and 117). The results in this table also indicate that the total number of inferences drawn from the rocket story in each condition (19, 18, 19, 21, 11, 39) was lower than the total number of inferences drawn from the radio story (30, 21, 28, 32, 23, 44). This may have been so because the rocket story was generally rather difficult for the Malawian children (see the inferences in relation to the familiarity of the story section below).
The results indicate that the participants drew more inferences when answering comprehension questions from the radio story than the rocket story. While the children verbalised only 49 inferences in English and 78 inferences in Chichewa about the rocket story, they verbalised 81 inferences in English and 97 inferences in Chichewa about the radio story (see
A comparison of the number of inferences drawn from the two stories.
Story title | Number of inferences (English) | Number of inferences (Chichewa) | Total number of inferences |
---|---|---|---|
‘The mole and the radio’ | 81 | 97 | 178 |
‘The mole and the rocket’ | 49 | 78 | 127 |
These results seem to reflect the fact that the rocket story is centred on the idea of travelling by rocket and crashing with the rocket on the seashore. In order to be able to return home, the mole then has to reassemble the rocket with the help of sea creatures. Overall the story seemed to appear alien to the children, who grew up in a landlocked country. We found in an earlier study that they struggled to retell this story (Hara
The results revealed that the participants were able to use a variety of knowledge bases in their inference-making processes to understand information that was not made explicit in the stimulus material (see
Inferences in relation to the children’s pre-existing knowledge bases for the radio story.
Condition | Elaborative inferences in relation to the question ‘what is the bee doing in the flower?’ |
||||
---|---|---|---|---|---|
Food: ‘Juice’, ‘food’, ‘sugar’ ‘water’, ‘honey’, ‘nectar’, ‘nutrients’ | Pollination | Habitation: ‘It is living in the flower’ | Other answers | Total | |
Condition 1 (a) | 9 | 1 | 0 | 1 | 11 |
Condition 2 | 12 | 1 | 2 | 2 | 17 |
Condition 4 | 8 | 2 | 1 | 3 | 14 |
Condition 1 (b) | 8 | 0 | 2 | 0 | 10 |
Condition 3 | 6 | 0 | 2 | 6 | 14 |
Condition 5 | 14 | 0 | 4 | 1 | 19 |
Use of knowledge transfer from other subjects for ‘the rocket story’.
Condition | Comprehension answers from the rocket story arranged according to different themes |
||||||
---|---|---|---|---|---|---|---|
Lack of fuel | Inability to control the rocket | Low batteries | Over speeding or over heating rocket | Intrusion of strong wind or water into the engine | Other answers | Total | |
Condition 1 (a) | 5 | 2 | 2 | 1 | 0 | 1 | 11 |
Condition 2 | 4 | 2 | 0 | 1 | 2 | 3 | 12 |
Condition 4 | 5 | 0 | 0 | 1 | 0 | 4 | 10 |
Condition 1 (b) | 2 | 5 | 0 | 0 | 0 | 2 | 9 |
Condition 3 | 2 | 4 | 0 | 0 | 1 | 5 | 12 |
Condition 5 | 1 | 2 | 0 | 2 | 2 | 4 | 11 |
With respect to the radio story, the participants were able to embellish their situation model of the story with information from a variety of school subjects, as they stated that a bee featured in the radio story was looking for ‘food’, ‘juice’, ‘sugary water’ in response to the question of what the bee was doing inside a flower. Four participants even used the technical term ‘pollination’, which confirms that the participants were able to tap into other knowledge sources, for example knowledge acquired in school subjects such as Biology (see
It is interesting that condition 3, where both the question and the children’s answers were formulated in Chichewa, yielded the highest number of alternative answers (‘other answers’ e.g. ‘the bee was playing inside the flowers’, ‘the bee was playing with his friends’, ‘the bee was eating the flowers’). This observation warrants further investigation.
Even though the rocket story was more difficult for the children to draw inferences from and to recall, they still managed to employ their knowledge base and were able to supplement the information that was explicitly provided in the film or audio recording with their own knowledge (see
In order to comprehend why the rocket had to crash-land, the children drew the following inferences: the rocket had no fuel; the mole was not able to control the rocket; the batteries were low; the rocket was over speeding or overheating; a strong wind or water may have entered the engine. Neither the film nor the audio recording included any explanation for the rocket’s engine failure.
The discussion in this section follows the sequence in which we present our findings in the previous section. First, we deliberate the relationship between inference-making and the children’s language constellation and relate these findings to the influence of the school type/socio-economic background. Subsequently we consider reasons why the participants drew more inferences from aural stories than visual stories and why there were more inferences from the radio story than the rocket story. In connection with the latter, we reflect on the relationship between cultural familiarity and comprehension skills. Lastly, this section discusses findings pertaining to the participants’ use of their knowledge bases in their inference-making processes.
Overall our participants drew more inferences when they answered comprehension questions in their L1 as compared to their L2. This is in line with findings that indicate that second language listeners have fewer working memory resources at their disposal than first language listeners, which makes listening in a second language an even more challenging task than first language listening (Horiba
Our results furthermore indicate important interactions between the children’s language constellations and the type of school that they attend, which in turn relates to their socio-economic background.
Starting with the latter aspect, we found that children who attended the private Kapita Primary School were generally able to draw more inferences than the children who attended the public Mphungu Primary School. This advantage of the children who attended a private school may be related to the fact that the school type corresponds to a higher socio-economic background. According to Hecht
Moreover, it is conceivable that the children who attended Kapita Primary School had the luxury of regular exposure to stories in the film medium both at home and at school, while many children attending the under-resourced Mphungu Primary School might not have enjoyed either of these privileges. Linebarger and Piotrowski (
Crucially, our results indicate that participants from Mphungu Primary School drew more inferences than participants from Kapita Primary School when answering comprehension questions in Chichewa, especially in condition 1 (b) (both stories) and condition 5 (i.e. the rocket story only), whereas participants from the private Kapita Primary School drew more inferences when answering comprehension questions in English (i.e. in all the assigned conditions). These results may be related back to the parental questionnaire, which indicated that the majority of children attending Mphungu Primary School (36/58) used Chichewa when narrating stories at home and had the ability to [re]tell stories in Chichewa (see
A further striking result of our investigation is that the participants drew more inferences when the mode of presentation was aural rather than visual (film). The children may have paid more attention to verbal text, a suggestion that is in line with Meringoff’s study (
Our results also confirm a positive relationship between cultural familiarity and reading or listening comprehension as indicated by Alptekin (
Finally, our results revealed that the children were to a certain extent able to use knowledge bases like, for instance, knowledge learned in a variety of school subjects such as science in their inference-making processes. This is in accordance with the finding that comprehenders make use of a variety of sources, for example knowledge about the language used in the text, general world knowledge, knowledge about the specific communication situation, domain-specific prior knowledge and personal experiences, to create situation models and hence fully understand stories (Meringoff
Inference-making in children has received a considerable amount of attention in the literature, and yet it is still understudied, in particular in bilingual as well as multilingual children and even more so in children who come from backgrounds that are different from mainstream US or European backgrounds that most of the current research is based on. The results of our study indicate that the pre-recorded audio recordings seem to have supported inference-making more than the corresponding wordless films. The results also indicate that the children from a private school (with English as a medium of instruction) drew more inferences than the children from a public school (where Chichewa is the medium of instruction). Furthermore, the results reveal that while the children were able to use knowledge transfer from a variety of knowledge bases to draw inferences, the inferencing process was impeded when the story content deviated too much from their own experience.
Our data indicate that we need to pay close attention to children’s language constellation and socio-economic background as well as their language socialisation, which includes their exposure to alternative narrative styles (for example, oral tradition and folktales). The influence of the latter is very poorly researched and needs further attention. Moreover, we need more research into the ‘transferability’ of inference-making abilities in one child across his or her different languages. We suspect that language attitudes may interfere with language competence in this domain.
We thank our participants and their parents, as well as Kapita Primary School and Mphungu Primary School for hosting our research. We also thank the audience of our talk presented at the joint annual conference of the Linguistics Society of Southern Africa, the Southern African Applied Linguistics Association and the South African Association for Language Teaching, held from 24 to 26 June 2015 at the North-West University, Potchefstroom, South Africa.
The authors declare that they have no financial or personal relationships that may have inappropriately influenced them in writing this article.
A.H. was a doctoral student at the University of KwaZulu-Natal with H.T. as A.H.’s supervisor, at the time the data were collected. A.H. tested the participants individually in a vacant room at their school. A.H. identified inferences from the participants’ answers; H.T. then verified these inferences. A.H. drafted the manuscript, which was rearranged and edited by H.T. A.H. compiled the results and presented them in tables. These results were checked and verified by H.T.
We are aware of the system dynamics theory, but our focus in this study was on the role played by inferences in the construction of a coherent mental representation of a text, hence our analysis of inferences.
The language choice for the parental questionnaires followed school policy: The head teacher at Kapita Primary School informed the first author that English was the medium of communication with the parents, while the head teacher at Mphungu Primary School confirmed that Chichewa was the medium of communication with the parents.
The films were chosen because they have enjoyed great popularity in many countries around the world since their creation.
In the original study, participants were required to recall the film or aural stories in accordance with the condition before answering the comprehension questions.
We found that all of the children were exposed to storytelling in the home. However, the children who attended Mphungu Primary School were predominantly exposed to storytelling in Chichewa, which indicates that they were exposed rather to oral, traditional storytelling (cf. footnote 7).
The proposed correlation between exposure to stories and performance in story recalls and comprehension tasks is further supported by numerous other pieces of research (e.g. Cruz de Quirós et al.
Obviously, not all parents from affluent homes engage in storytelling sessions with their children. It is possible for an affluent home to be headed by a parent or parents who do not have time to read stories to their children.
This hypothesis is partially based on the fact that there is a paucity of reading material in Chichewa and we hence assume that the children and their families engaged in the narration of Chichewa folk tales. We found evidence for this assumption in our earlier investigations of the same children (Hara