In this article, interpretation and analysis of the integrated data sets were guided by a general perspective on learning that emphasises integration of thinking (cognition), feeling (affect), and acting (conation and associated behaviour). According to this perspective, desirable learning proceeds with the learner both willing and able to take responsibility and control over personal learning practices. Particular themes are developed to structure data interpretation in terms of the learning constructs discussed below. These themes centre initially upon aspects of thinking then extend to the affective and conative domains.

Ten terms that structure discussion to follow are shown in Figure 1. These terms form the basis of the learning model to be considered later.

Figure 1: Ten Learning Terms

In Figure 1, the terms are arranged somewhat sequentially in terms of learning approach, progress and outcomes (Baird, 1998). These terms cover cognitive, affective, and conative aspects of learning. I will now briefly review these aspects, argue for their effective integration in attempts to enhance learning quality, and then explain the place of ignorance in the figure.

Excluding ignorance, the constructs start with perceptions, motivation, and attitudes/emotions, terms that are clustered together to emphasise the importance of (cognitive) perceptions for all that occurs next, and the influence of these perceptions (whether accurate or inaccurate) on various feelings. The nature of perceptions and associated affective elements will determine the extent to which the learner is positively, productively challenged by what is to be done. I have defined and discussed perceived challenge elsewhere (Baird, 1994; Baird and Penna, 1997) as not simply a measure of extent of thinking (cognitive demand), but also the product of extent of interest and motivation (affect). If interest and motivation are high, cognitive demand will be viewed positively, with the likelihood of a sense of productive challenge, leading to the next step in the process. If a sense of positive challenge exists, the learner will be disposed to engage in active reflection and associated action to set worthwhile learning goals and to achieve these goals. In turn, these processes involve metacognition and volition. I define metacognition as having three components: metacognitive knowledge (knowledge of the nature of learning, effective learning techniques, and personal learning characteristics); metacognitive awareness (of task nature and progress - essentially knowing what you are doing and why you are doing it) and metacognitive control (making productive decisions about approach, progress and outcomes) (e.g. Baird, 1991). Volition connotes strength of will and a sense of striving; it is associated with such attributes as diligence, perseverance and resourcefulness (e.g. Corno, 1993). Volition essentially means determination to carry out decisions made, by controlling unhelpful emotions and resisting distractions. Both metacognitive and volitional control are executive control mechanisms, with volitional control (meta-emotional and meta-motivational control) providing the necessary affective focuses to complement the cognitive focus of metacognitive control. As indicated in Figure 1, the desirable ultimate outcomes of learning are both understanding and enjoyment.

The arrangement of terms in Figure 1 mirrors recent attempts by others to view learning in a way that emphasises its purposive character, where such things as perceptions, aspirations, feelings about self (e.g of competence and confidence), decisions, learning goals, and behaviours are brought together to inform and illuminate learning actions and the meanings that underlie such actions. Snow, Corno and Jackson (1996) present a provisional taxonomy of individual difference constructs that links cognition (procedural and declarative knowledge), affect (or affection: temperament and emotion) and conation (motivation and volition). Research into the increasingly popular areas of "learning to learn" (e.g. Candy, 1990) and of self-regulated learning (e.g. Boekaerts, 1997, 1998; Zimmerman, 1998; Zimmerman and Schunk, 1989) emphasise interactions between perceptions, learning goals, metacognition and volition. For instance, Zimmerman (1998, p. 4) presents learning as a "self-fulfilling cycle" that encompasses forethought (e.g. goal setting, strategic planning, intrinsic interest), performance/volitional control (e.g. attention focussing, self-monitoring) and self-reflection (e.g. self-evaluation, attribution, adaptivity). Boekaerts (1997) distinguishes between cognitive self-regulation and motivational self-regulation in a six-component model of self-regulated learning (p. 164). From a different perspective, Boekaerts (1998) considers self-regulated learning according an action hierarchy involving multiple cognitively- affectively- and volitionally-related goals at various levels of abstraction (from the most general level, self-concept, through to more specific levels of principle control, action programs, and scripts - p. 17)

It may seems strange to the reader that ignorance was the first term listed in Figure 1. It was placed there because ignorance is both a vitally important and a common feature of all learning. Ignorance - not knowing something or how to do something - is a common condition in daily life and is the basis for new learning.

As important as ignorance, however, is whether or not you are aware of it. Sometimes you realise that you don't know something , sometimes you don't. Knowing that you don't know or can't do something is prerequisite to effective learning. Indeed, this situation means that the learner has achieved metacognitive awareness, which is necessary for adequate metacognitive control over learning. Alternatively, you may not know that you don't know or can't do something. In this situation, you are metacognitively unaware - a term that has been entitled secondary ignorance (e.g. Brown, 1978).

An unfortunate aspect of the first situation above is that many metacognitively aware students may be disinclined to act to convert their ignorance into understanding. In this case, the problem is affective rather than cognitive, and the outcome is inaction. A reason for not acting on recognised ignorance is to mistakenly relate ignorance with stupidity, and consequently to make the decision to not reveal ignorance for fear of appearing foolish. Not taking the risk to betray personal ignorance would be a tactic especially favoured by those students who lack confidence in their ability as learners (e.g. Miserandino, 1996) and exhibit ego-involvement rather than task-involvement in their achievement motivation (e.g. Nicholls, 1983; Spaulding, 1992).

The other undesirable learning situation, where the learner is metacognitively unaware of the nature and extent of personal cognitive ignorance, has all of cognitive, affective, and volitional aspects that act together to directly diminish a person's willingness and ability to achieve desirable learning outcomes. One type of metacognitive unawareness is when you don't realise that you don't know or can't do something. This unawareness may arise from inaccuracy - for instance, when you think you know what you are doing when in fact you don't. This situation is often involves the setting of unproductive learning goals. A similar but different type of metacognitive unawareness is if you don't know whether you do or don't know something. Here, motivation would be expected to be low and emotions negative, because you don't know what to do next.

In Figure 2, the ten constructs in Figure 1 are subsumed within an overall model of learning that formed the frame for interpretation of data arising from the integrated data sets.

Figure 2: Learning Model

Figure 2 emphasises interaction in learning between the constructs shown in the large triangles in the centre of the figure. In one triangle, perceptions (cognition) and motivation and emotions/feelings (affect) generate a level of metacognitive awareness of self, task and context conditions. These constructs influence, and are influenced by, learning decisions and associated behaviours, evidenced by levels of metacognitive and volitional control. Desirable behaviours comprise actions taken in response to focussed, purposeful reflection (usually mediated by directed evaluative question-asking). These reflective behaviours and their related levels of metacognitive and volitional control in turn will determine quality of the overall learning outcomes, as shown on the right of the figure. The nature of desirable self-regulated learning, as discussed above, is now considered in terms of contrasting types of interaction between the constructs in the two triangles.

In an undesirable type of learning that I entitle Perceptions-Driven Learning (PDL - Figure 3) the learner approaches the learning task reactively, with few self-generated reflective behaviours. In PDL, task, context, and self conditions - that include nature and level of cognitive ignorance - are inadequately or imperfectly diagnosed. Lack of active reflection results in faulty or unproductive perceptions that, in turn may be associated with negative emotions regarding what is to be done, with resultant low level of perceived personal challenge. Because metacognitive awareness is low, metacognitive control is limited, and this situation may result in unproductive decision-making regarding learning approach. Subsequent faulty metacognitive and volitional control over learning progress precludes success in learning.

Figure 3: Perceptions-Driven Learning

Reflection-Driven Learning (RDL - Figure 4) is now presented as an alternative to Perceptions-Driven Learning.

Figure 4: Reflection-Driven Learning

In RDL, the learner has habituated a process of purposeful enquiry, involving active reflection and associated action. The learner's manner of engagement in the learning task is to actively interrogate its nature, characteristics and purposes, and to determine personal attributes for successful completion. Accordingly, the learner determines what needs to be done, and what he or she does and does not now know - thus achieving metacognitive awareness. A key desired outcome in the learning model shown in Figure 4 is understanding. Understanding is achieved by appropriate reflection upon aspects of cognitive ignorance and action to remedy such ignorance. In Australia, the Project for Enhancing Effective Learning (PEEL - Baird and Mitchell, 1986; Baird and Northfield, 1992) continues as a significant school-based project to help students to become more willing and able to engage in such reflection and associated action, to take greater responsibility for, and control over, their own learning and thus, to convert ignorance into understanding.

Below, I provide a brief summary of support arising from this study for the assertions made above. Findings that inform the theoretical perspectives summarised above are considered according to four related themes.

Learning decisions and teaching behaviours;

Learning decisions, perceptions and motivation;

Learning decisions: deficient perceptions and metacognitive unawareness;

Learning behaviours and teacher perceptions.

As mentioned at the start of this article, themes and supporting statements are made without presentation of empirial data. Findings arising from data analysis are simply listed in dot point form under theme headings. Many of the themes and findings are elaborated in the presentation.

LEARNING DECISIONS AND TEACHING BEHAVIOURS

• Repeatedly, teaching behaviours had a profound influence on decisions made by students in their learning. Some of these decisions were productive; some were not.

• There was some evidence of the teacher grounding teaching practice in educational theory or productive prior personal experience.

• There were also various attempts by teachers to enhance students' interest and understanding by relating work to the world and to lived experiences:

• Alternatively, many teacher behaviours militated against students approaching their learning systematically and setting worthwhile learning goals. Three main types of behaviour are:

• Vagueness or ambiguity in classroom instructions. This problem was common; there were at least fourteen episodes related to it for the three lessons from Mrs. B. alone.

• Requiring students to undergo mindless compliance or "guess what's in the teacher's mind":

• Compounding the difficulties in learning that arose from these type of teacher behaviour, teachers sometimes either lacked knowledge regarding some key content areas, or in some cases had inadequately prepared for the lesson.

• Perhaps as one result of such behaviours, the teacher occasionally acknowledged only a limited idea of lesson learning outcomes:

• Students' willingness and ability to make productive learning decisions was influenced clearly by factors such as those above.

• In addition, the nature of students' learning was influenced by their perceptions of the teacher, their role as learner, and the nature of the learning activity.

• Students' perceptions of the teacher, or of their role as learner, directly influenced their learning decisions and behaviours.

• There were many instances of inferred and admitted cognitive ignorance. Methods of dealing with such ignorance varied considerably, to generate desirable and undesirable learning.

• On many occasions, students clearly demonstrated that they were metacognitively aware of personal ignorance.

• Cognitive ignorance (and confusion with which it was associated) had a powerful influence on decisions made and opportunities to effect metacognitive control over learning:

• Alternatively, there were instances of students dealing more effectively with acknowledged ignorance, through application of Good Learning Behaviours (GLBs - Baird, 1986) to pursue useful learning goals, and reaching a satisfactory learning outcome.

• On some occasions, students' decisions to ask other students or the teacher to resolve instances of personal ignorance resulted in a negative outcome.

• On other occasions, however, when students demonstrated an awareness of their cognitive ignorance and then, through reflective thinking, sought knowledge from colleagues or the teacher, the result was positive.

• A student's willingness and ability to undergo the thinking, feeling and acting necessary to transform perceptual and conceptual inadequacies and inaccuracies such as those above into understanding are influenced fundamentally by the social situation. This social situation involves interpersonal interactions with other students and with the teacher.

• Data relevant to this section concentrate upon the thoughts, feelings and actions of one Year 9 student, student J, and the teacher's perceptions of, and interactions with, her.

• Student J had motivational and volitional limitations to her learning, brought about by her perceptions of the unchallenging nature of the work, and her feeling that the teacher "picked on her".

• In comparison with many instances of poor learning were many clear instances of episodes where she adopted a focussed, reflective approach to her learning, especially when she found the task interesting.

• At interview, there were many occasions when she demonstrated a clear lack of metacognitive awareness of aspects of the teaching, and its effect to limit her motivation, commitment, and understanding:

• A most important aspect of the influence of social interactions for nature and quality of learning centred on teacher's perceptions of Student J. At interview, the teacher repeatedly expressed attitudes and beliefs that denigrated Student J and her learning. These negative attitudes and beliefs seemed to be sufficiently powerful that they precluded the teacher making more reflective evaluation of Student J's learning behaviours; Student J was aware of this negativity.

• When Student J viewed the classroom videotape, she came to a realisation about how her own behaviours could easily be misinterpreted as lack of attention by the teacher. Perceptively, she decided to try to amend such behaviours in the future.

Classroom data are complex and multiple. Worthwhile, trustworthy ethnographic research needs to present selected data in a way that invites interpretation of observed actions in terms of possible underlying reasons or purposes. Well-grounded interpretive research requires some credible and relevant referential bases that allow isolated episodes or events to be organised, considered and disseminated coherently. The data obtained in this study provided evidence for personal and interpersonal actions and learning conditions that either help or hinder conversion of cognitive ignorance to understanding. The data also provides support for the inextricable contributions of all of cognition, affect and conation in determining learning approach, progress and outcomes. Above all, perceptions of self, task and context were demonstrated to be central to levels of motivation, perseverance, and informed progression through metacognitive and volitional control.

Teaching and learning are social activities. Consistent with my prior research in this area, it seems that both teacher and students will be stimulated to think, feel, and act to transform personal ignorance into understanding when each person perceives personal productive Challenge in the learning tasks, and also feels that the classroom climate allows for positive interpersonal interactions in a supportive atmosphere. Crucial to generation of such an atmosphere is first that "it is O.K. not to know" and second that the lesson activities are such that students feel that "I know that I don't know, and I want to find out".