Published as: Wegerif, R. (2000) Learning to Talk: Talking
to learn. Teaching
Thinking, Vol 1, issue 1.
New research at the Open University has demonstrated a link between teaching the ground-rules of democratic debate and measurable improvements in the reasoning test scores of children. The research team, consisting of Neil Mercer, Lyn Dawes and myself, worked closely with local primary teachers to develop a set of 'Talk Lessons' to promote listening with respect, responding to challenges with reasons, encouraging all in the group to give their views and trying to reach agreement before moving on. We also developed a number of computer-based activities that, used with the ‘Talk Lessons’ gave small groups of children an opportunity to discuss problems and issues together. Through careful evaluation we have shown that the use of this kind of talk increases the effectiveness of group work, improves learning in curriculum areas and even leads to increased results for children on tests of reasoning.
Our research was designed to explore a ‘dialogical’ model of how children learn to reason. On the dialogical model reasoning is not a property of the individual child, like his height or hair colour, but a property of conversations. Learning to reason involves ‘internalising’ the rules of good conversations. In other words the theory suggests that effective individual thinking is rather like having a good conversation inside your own head. If this theory is correct then it follows that teaching children the ground-rules of effective dialogue with others should lead them to be better at individual reasoning tests. The kind of reasoning test which has been widely taken to measure that aspect of intelligence which is most individual because most 'context-free' are non-verbal tests such as the Raven's Progressive matrices test. The Raven's test consists of a series of geometric shapes where the children have to discover the pattern in order to continue the series. One of these problems is illustrated in Figure 1. The test of the theory which we devised was this: if we can we improve individual results on Raven's tests by teaching them only the social ground-rules that promote dialogical reasoning then this will indicate that the dialogical model of learning thinking works. To implement this test we first had to specify more carefully what we meant by good dialogical reasoning.
To test the idea of dialogical reason we needed a practical set of social ground rules that we could teach to children in classrooms. We developed the following rules partly influenced by a survey of the literature on effective collaborative learning, partly influenced by the philosophy of rationality but mainly based on our experience in classrooms working closely with teachers. From these different sources seven pragmatic ground rules were put forward:
1 all relevant information is shared
2 the group seeks to reach agreement
3 the group takes responsibility for decisions
4 reasons are expected
5 challenges are acceptable
6 alternatives are discussed before a decision is taken
7 all in the group are encouraged to speak by other group members
Once we had decided on our ground-rules to support reasoning we designed ten lessons to teach these rules including two activities around software. This task mainly fell to Lyn Dawes who has twenty-three years experience as a primary school teacher. It would take too long here to explain in detail how the ground-rules were taught except to say that each lesson involved a combination of small group and whole class work. In one of the early lessons the class teacher began by asking the children to discuss a story in groups of three, the discussion was guided by a worksheet. After this activity the children were brought together into a whole class discussion and asked to feed-back not only what they had thought about the story but also what they thought about how they had talked about it together. The children were then led to suggest rules for working together more effectively and these rules were written on the board for more discussion. Out of this discussion, with a bit of leading by the teacher, a set of ‘ground rules for talk’ was produced. This list of ground-rules, written in large letters, was then displayed on the wall of the classrooms. In all succeeding talk lessons these rules were referred to as ‘our rules’ by the teacher and sometimes, also by the children. I hope that some of the other activities we devised will be the subject of a separate article to appear in a later edition of this journal.
In two separate studies classes of children were taught the ground-rules of reasoning. The impact of this teaching on their thinking was looked at by giving them two versions of the Raven's tests one of which they were asked to talk around in groups of three and the other of which they had to do alone. These tests were given to the children before the educational programme began and again afterwards so that we could compare their reasoning together on the same problems. In both studies the groups did much better afterwards than they had done beforehand. Here is an example of a group talking around a Raven's problem. Before they are given our set of lessons they get the problem wrong but afterwards they get it right. It is interesting to consider why this is. I suggest that you look at the problem before you read the transcripts.
(In the following transcripts all names have been changed and comments based on the video recording are in parentheses)
The initial talk of
John, Elaine and Graham around problem A11
Elaine: No, because it will come along like that (Elaine rings 5 as answer for A11)
The talk of the same
group on the same problem after our teaching programme
John: Number 5
Graham: I think it’s number 2
John: No, it’s out, that goes out look
Graham: Yeh but as it comes in it goes this
Elaine: Now we’re talking about this bit so it can’t be number 2 it’s that one
Elaine: It’s that one it’s that one
Graham: Yeh 'cos look
Elaine: 4
Graham: I agree with 4
(John nods his assent and Elaine rings 4 as answer for A11)
This is one of the earlier problems in the Raven's series and one that most groups got right the first time. If you look just at the darker lines running vertically you might think that the answer is number five because that continues the pattern for these lines. This seems to be what happens the first time that the group of children looks at the problem. Elaine, who is perhaps the dominant member of the group, ringed number five, which is the wrong answer. She did not pause to try to get agreement. The fact that she said ‘because’ indicated that she was responding to someone else's suggestion made through pointing at one of the pictures. In the first test the children were rushing through the problems without much discussion.
After the series of ten lessons in the ground rules of reasoning given over ten weeks the children take more time over the problems. Reconstructing how they solve the problem together illustrates why three heads are sometimes better than one, or even than two. First, as before, the pattern of the dark lines is spotted and John puts forward number five as the answer because it fits this pattern. But this answer is only a suggestion. Graham puts forward number two as the answer because he is looking at the horizontal pattern of the lighter lines. John explains with a mixture of words and pointing that the vertical black lines have to go out. Graham in turn explains that it cannot be number five because the light lines have to go in. Each boy has adopted one perspective; John takes the side of the dark lines, Graham that of the light lines. Each can see enough to refute the position of the other but this does not produce the solution. Elaine then, helped by the discussion and the reasons given, comes up with the answer which combines the dark lines going out with the light lines going in, that is number four. Both Graham and John can see that she is right according to the same arguments that they put forward just before.
These children found a solution to A11 through exploring different alternatives, responding to challenges with reasons and trying to reach agreement before moving on. This illustrates how applying the ground-rules of reasoning which were taught to them helped this group to reason together better about raven's test problems.
It is perhaps not very surprising that children who were taught the ground-rules of democratic debate did better at solving raven's tests in groups than classes of children who had not been taught to use these ground-rules. What is perhaps more surprising is that these same children also did better working as individuals. In the last study we did on the effects of teaching the ground-rules of reasoning we worked with three teachers to teach these ground-rules in three classes of mixed year four and five children. We also gave the same tests about 10 weeks apart in three similar classes in different schools to act as a ‘control’. The results of these tests showed not only that the groups in the classes which we taught improved their scores on the test by about 10% but also that the same children working on their own using a similar test increased their scores more than the control children. Analysis showed that this increase in individual scores was statistically significant. (Fuller figures are given in an article in the British Educational Research Journal, Mercer, Wegerif and Dawes, 1999). This is an exciting result because it confirms the hypothesis we started out with, that so called ‘non-verbal’ reasoning ability which was supposed by some to be the most ‘pure’ measure of general intelligence can be improved by teaching children ground-rules for good discussions. This result offers support for the theory that individual thinking skills are internalised talking skills. In the next section I will expand a little on what this might mean for the way that we think about reason.
Mathematical images of reason are widespread in the psychology of education. Piaget's theory of how children learn to think was based on the idea that they had internal logical or mathematical structures that developed through interaction with the world around them. A variant on this essentially mathematical picture of thought and thinking skills can be found in the computer models used by cognitive psychology. Thinking seen on the model of mathematics or computing can justly be called 'monological' because it assumes a single logic, a single voice and a single true vision. Thinking seen on the model of dialogue, a ‘dialogical’ model of thinking, assumes that there are always at least two voices involved in constructing meaning. While monological models of thinking focus on abstract structures supposedly located within individual brains, dialogical models see thinking as something that occurs in conversations between different voices and perspectives. According to the dialogical model, without differences in perspective there would be no dialogue and therefore no reasoning.
Some contrasts between monological and dialogical models of thinking are given in table 1
|
Monological |
Dialogical |
|
one voice |
many voices |
|
self-identity |
difference |
|
representation (in a structure) |
relationship (with a person) |
|
static |
dynamic |
|
de-contextualised |
contextualised |
|
predictable |
unpredictable |
|
univocal |
equivocal or ambiguous |
|
closed system |
open system |
The idea of dialogical reason is not a rejection of reason but a new way to understand what reason really is. Thinking about thinking on the model of dialogue may do away with notions of absolute truth and even perhaps with the importance of fixed rules of logic but this does not mean that anything goes. When we join with others in a free and open discussion in order to try to reach a shared decision real thinking goes on, or it should do if participants enter into the spirit of the dialogue and are prepared to learn from it. This is why there is a connection between good citizenship and good thinking. On this dialogical model of reason, to be a reasonable person does not mean that you are an expert at applying the rules of logic or mathematics but that you know how to participate constructively in learning conversations. A key feature of being reasonable is that you are able to listen to others and to change your own position if their reasons are better than yours.
The structuring ground rules for effective learning conversations are more like moral virtues than they like the rules of logic. They are rules such as encouraging the expression of alternative views, listening with respect to opinions wherever they come from, being prepared to change beliefs and always seeking to persuade others with reasons that they will accept rather than trying to coerce, bribe or brow-beat them. The exact nature of the ground-rules that work best to create new understandings and new knowledge in different contexts will always be open to challenge and debate but we all know intuitively that there are such ground-rules.
Raven’s tests were not the whole focus of the study. We videotaped children talking together around a range of problems in different curriculum areas. In particular we designed and made software to support discussion to help learn about friction which is part of the science curriculum and to help learn about morality in the area of Personal and Social Education that may now be called Citizenship. We found that computers were a good way of stimulating group discussion. Computers can prompt a discussion like a teacher, but unlike many teachers they are not judgmental and are infinitely patient. Once children had gone through our ‘Talk Lessons’ we found that when the computer asked them a question they were able to sit back from the screen and discuss it together before reaching a shared decision about what to do next (i.e. what button to click next with the mouse)
In the following extract a group of three year five children are working together around a computer which is taking them through a story about the moral dilemma faced by a girl called Kate. Kate has promised her friend Robert not to tell his secret, which is, she then learns, that he stole a box of chocolates. What should Kate do now? Before the lessons in the ground rules of reasoning all the groups of year 5 children (age 8 and 9) to whom we gave this software answered ‘Tells her parents’ with little pause for thought. It seems that they thought this must be the answer that the teachers want them to give. After the educational programme the children talked together much more and appeared to really think about the issue. The following transcript extract shows how Kylie, Tom and Gary answered the computer prompt.
(== Continuous talk. {Overlapping talk. / turn completed by another speaker)
Kylie: What do you think?
Tom: What do you think?
Kylie: No
Tom: Why? I don't agree
Kylie: We said why
Kylie: I think that one as well do you?
Kylie: I think I'm I think even though he is her friend because he's stealing she should still tell her parents and her parents might give her the money and she she might be able to go to the shop and give them the money
Tom: I think um/
Tom: {I think that he should go and tell his mother /…
Kylie: …/even though she has promised
Tom: Because he's …well you shouldn't break a promise really should you
Kylie: If he steals … If you know he's stolen if she don't tell her parents then he will be getting away with it
[3 second pause]
Tom: OK then
Kylie: Ain't worth it is it?
Tom: Tells her parents
This is not a perfect debate by any means but these children were not perfect children. Their school had a catchment area with social problems and a high proportion of the children had ‘statements’ for learning difficulties. This particular group had found it very difficult to work together at the beginning of our ‘talking lessons’. In this transcript it is noticeable that the children are not afraid to challenge each other and that they respond to challenges with reasons and reflection. This is very different from their behaviour before the ‘Talk Lessons’ when challenges led to aggression. Following the ground-rules apperas to help them to think about the issues of stealing and friendship more deeply.
Achieving improved results on tests of abstract reasoning might seem a bit irrelevant to ordinary classroom education. The interest of these results is that they establish a connection between group thinking, of the kind found in learning conversations, and individual thinking ability of the kind measured by tests. These results show that teaching children to talk together better is a way to teach them to think better even when they are working on their own. And although the Raven’s tests might seem very abstract the scores that children achieve on these tests correlate well with their later academic achievement. In other words teaching children a way of reasoning which helps to solve Raven’s tests is also probably teaching them a way of reasoning which is more generally useful in education.
However I think that our research goes further than showing how education can be made more effective within the current curriculum. It does that certainly but it also suggests the basis for reform of the curriculum. Our approach is to give children the communication skills and opportunities that they need to participate in shared reasoning. The ideal of dialogical reason is of a free an open discussion of all relevant alternatives where conclusions are reached by consensus through the quality of arguments. This ideal is intimately connected with the ideal of democracy. The kind of conversations we encouraged in the classroom are the same kind of conversations which are supposed to take place wherever important decisions are taken in a democratic society. Teaching children this way of communicating is not simply teaching them thinking skills so that they can improve educational test scores, it is also a way of giving them the possibility of becoming full participants in the conversations that shape their lives.
Dawes, L.
(1997). Teaching Talk. In Computers and Talk in the Primary Classroom,
edited by R. Wegerif and P. Scrimshaw. Clevedon: Multi-lingual Matters.
Mercer, N.
(1995) The Guided Construction of
Knowledge: talk amongst teachers and learners. Clevedon: Multilingual
Matters.
Mercer,
N., Wegerif, R. and Dawes, L. (1999) Children's talk and the development of
reasoning in the classroom, British
Educational Research Journal. Vol. 25, no 1, p 95-111
Wegerif, R., Mercer, N., & Dawes, L. (1998). Software design to support discussion in the primary classroom. Journal of Computer Assisted Learning, Vol. 14, no 3 199-211.
http://www.thinkingtogether.org.uk