Reports from the Department of Education and Educational Research, Göteborg University, Sweden, No. 1996:05
INTRODUCTION
For Nordic education researchers, phenomenography is generally connected to work at the Department of Education and Educational Research in Gothenburg, Sweden, led by Ference Marton and his associates. However, the research from which the term phenomenography derives has been carried out at there since the early 1970:ies (cf. Marton, 1988) with some interesting variations and similarities evident. This paper comments on some of these and their relations to phenomenology, the branch of philosophy to which some phenomenographers have looked for support in recent years. It also questions how reasonable this link is.
Not much material on the theoretical underpinnings of Gothenburg phenomenography is internationally available from which to develop a picture of what it really is or involves. In the coming pages we will make some attempts to unearth the foundations of Gothenburg phenomenography, sketch a background to its emergence and explore its potential links with phenomenology. That is, we will try to examine Gothenburg phenomenology as a research posture (cf. Wolcott, 1992) in relation to the research postures of empirical forms of phenomenology.
Research postures, as Wolcott (1992) points out, carry with them suggestions concerning suitable kinds of methods of research which may be transformed or adapted under specific circumstances within the parameters of reasonable variation carried by the posture in specific research situations. The variations carried in phenomenography are explored by us in this paper, where our concern will be for variations in research design, their implications and injunctions. That is, how phenomenographers define, investigate and portray their research object.
Our conclusions are that a considerable amount of phenomenographic research has been carried out and some striking varieties of work are included, and that as yet there have been relatively few analyses of the approach. In fact, except for Marton's article in the journal Instructional Science (Marton, 1981), the Gothenburg version of phenomenography is only usually accounted for in anthologies (e.g. Marton, 1988) or more regular textbooks (e.g. Larsson, 1986; Uljens, 1989). Marton's text in Instructional Science was until recently the only regular scientific text from the Gothenburg group covering the theoretical fundamentals of phenomenography. Our paper seeks to provide ideas about avenues of exploration with regard to legitimate and problematic variation in phenomenographic research.
THE ORIGINS OF PHENOMENOGRAPHY
The phenomenographic research interest at Gothenburg developed firstly in the TIPS-project (Tillämpad inlärningspsykologi och studiefärdighet: applied learning psychology and study skills) at the Department of Education and Educational Research, and subsequently through the research activities of a group at that department which became known as the "INOM-group" (INlärning och OMvärldsuppfattning: learning, and conceptions of the world around us), a group whose concern was for both what and how people learn from their world. At this time there was no label "phenomenography" to denote this research interest, nor were the researchers particularly concerned with phenomenology. Research interests were mainly emprical, for study skills and successful learning as indicated by the "quality" of learning outcomes (cf. Marton, Dahlgren, Svensson & Säljö, 1977).
These researchers felt they had made one really "key discovery" in their work. This can be summarised as follows, and became foundational for further research: that understandings of whatever phenomenon or situation we take will, in a sufficiently large population or sample of people, vary in a limited number of qualitatively different ways which are crucial for the quality of subsequent learning and also its outcomes (cf. Marton, 1995; Marton & Booth, 1995). Two important strands of research developed from these foundations. One was concerned with ways of experiencing something, its object being to disclose the variation of ways of experiencing, the second took as its intention to describe phenomena in the world as others see them (Marton & Booth, 1995). The latter more so than the former became associated with phenomenography, indeed the term was "coined" by Ference Marton and accepted by his associates to denote it.
EARLIER CONCEPTIONS OF PHENOMENOGRAPHY
The word phenomenography has its etymological roots in Greek phainomenon (appearance) and graphein (description), rendering phenomenography, a description of appearances (Bengtsson, 1993). Other roots have been traced to modern phenomenology, founded by the German philosopher Edmund Husserl in his Logical Investigations in 1900-01 and developed into a movement by scholars like Heidegger, Sartre, Merleau-Ponty, Gadamer and Schütz (Bengtsson, 1993).
According to our research, the term phenomenography first appeared in research texts in 1954, in an article about phenomenology and existential analysis by Ulrich Sonneman (Sonneman, 1954), in which Sonneman used the term to distinguish Jasper's psycho-pathological research from other phenomenological works, particularly those of existential phenomenology.* Sonneman felt Jasper's work would be better termed as phenomenography because of its essentially descriptive character and gave phenomenography the following definition:
phenomenography, a descriptive recording of immediate subjective experience as reported, for example, by a person under psychiatric examination, without questioning the share in such a communication of the ego. (Sonnemann, 1954, p.344)
Jacob Needleman also touched upon the meaning of phenomenography when writing a critical introduction to his translations of the selected papers of Ludwig Binswanger, in which Binswanger's existential psycho-analysis is presented (Needleman, 1963). Needleman devoted a little more interest than Sonneman to what phenomenography might be about, and also ascribed it the quality of being a "good-for-nothing brother of phenomenology" (Needleman, 1963, p.37). Needleman was not unreservedly positive in his comments on phenomenography:
Phenomenography comes into being when the lesser circle grows uncontrollably, when the criterion for primitive fact becomes infinitely lax. In such a case, the possibility of the greater circle even emerging at all disappears, and with it disappears the possibility of system even while, at the same time, the attempt at system is being made. The result is encyclopedic, not explanatory. (Needleman, 1963, p.37)
Without developing the analysis of the above works further, in the views of both Sonnemann and Neeedleman, phenomenography emanates from psycho-analysis and was inferior to existential phenomenology.
Gothenburg phenomenography
The meaning Gothenburg phenomenographers give to the term phenomenography is not always clearly spelled out. We feel that there are several possible reasons for this, but that one of these is that phenomenography as practiced in Gothenburg has its origins in pragmatic empirical research where the need to spell out ones philosophical commitments was not strongly felt. Indeed, it is only relatively recently that phenomenographers have begun to seek a philosophical foundation. This has often been in phenomenology and has occurred in response to criticism from outside the Gothenburg bastion (cf. Uljens, 1993).
In the Swedish National Encyclopedia, phenomenography is presented in line with the Gothenburg tradition as phenomenological education in the following way:
phenomenography, /Š/ aims at identifying and describing the qualitatively different ways people conceive of different phenomena. (Nationalencyklopedin, vol. 6, p.183; our translation)
Marton (1992, p.253) describes phenomenography as a research approach for describing the limited number of qualitatively different ways in which a phenomenon is experienced, conceptualized, or understood, based on an analysis of accounts of experiences as they are formed in descriptions produced in research with other people. As we will go on later to elaborate, usually these accounts are produced exclusively through interviews, although there are some exceptions (e.g. Lybeck, 1981; Ahlberg, 1992; Booth, 1992, Alexandersson, 1994).
As can be suspected, like other forms of research the phenomenographic venture is problematic. In phenomenography many of these problems circulate around the claim phenomenographers make to be able to describe others experiences of a phenomenon as this implies living the experience of a phenomenon vicariously, through the eyes of another, by stepping back from ones own experience and using it only to illuminate the ways in which others state an understanding for something. This is a second order reflection over emprical depictions of the nature of an experience of something specific. Phenomenography has a characteristic form of empirical production which reflects this.
The problems of second order research have however often been ignored by phenomenographers themselves. Some of these acan be related to traditional problems of content and construct validity. More critically we feel though is the lack of critical reflection over the production relations of phenomenographic work which we feel phenomenographers also evidence: fundamentally, a lack of reflection concerning how the data and the findings of phenomenography reflect the understandings and experiences of the subject(s) of the research. This lack of reflection crosses both epistemological, methodological and ontological reflexivity. Both the data, the constructs or findings and even the object of phenomenographic research itself may be reflections of the researchers own ideas and/or products of interaction in the empirically productive situation (e.g. interview) or reflections of the general interests of a particular historical moment in qualitative research because of this (cf. Francis, 1993; Säljö, 1994; Marcus, 1994; Bowden, 1995).
These possibilities have not really been penetrated in phenomenography. Rather, in phenomenography the idea that data are descriptions of the internal relation between persons and things seems often to be unproblematically accepted. That is, the original and founding point of departure, that whatever situation we meet can be and is understood in ways which can be reduced to a limited number of qualitatively different ways of experiencing, is not put at issue.
Lack of reflexivity
Most Gothenburg phenomenographers pay little attention to the problem of reflexivity in their work. However, some have recently attempted to articulate a view of human experience to compensate to some degree this lack of concern. Experience is now defined as non-dualistic in Gothenburg phenomenography (cf. Marton, 1995; Marton & Booth, 1995) with the following implications: (a) That in descriptions of their ways of experiencing something, people say something about both the person and the field or domain of their experience. This is because (b) they are fused and something of the subject and of the object is in any relation of this kind declared in this way (Marton & Booth, 1995).
The question is whether this is enough. As we see it, the definition reflects a rationalist analysis of what phenomenography can reasonably comment on, which is itself rooted in an attribution of truth to phenomenography rather than a reflection over the production relations of phenomenographic work, their injunctions and their consequences for definitions of what falls within the framework of truth possibility in this tradition (cf. Rabinow, 1986). That is, qualitative differences in person-world relations are rationally defined as existing as distinctions between what can be called to mind about the thing(s) which consciousness is directed toward by a person, and are a priori attributed to say something about how these elements are related to each-other (structurally, sequentially, discriminantly).
The non-dualist assumption attributes a particular ontological status to the roots of differences in ways of saying something, which are said to be in a mind-world relation. By simply accepting this, the differences can be categorised as representing something other than just what people say, to "save" the results of phenomenographic analysis from criticism. Even Marton and Booth (1995) indicate a danger in accepting the non-dualist assertion too lightly:
The way in which a person experiences a phenomenon /.../ constitutes one facet of the phenomenon, seen from that persons perspective, with that persons biography as background. When the researcher describes the differing ways of experiencing a phenomenon /.../ (s/he) is (only) describing the phenomenon /.../ partially from the reports or inferences of the subjects, and it is this partial constitution of the phenomenon which is the researchers description. (Marton & Booth, 1995: p.126)
The above cited work is not yet publicly available and the warning carried by it has not been heeded by most phenomenographers. As we see it most phenomenographers simply accept their findings as genuinely denoting a map over the field of existing person world relationships regarding a particular phenomenon. This is clear by recourse to Sandberg (1995, p.155) and Bowden (1995, p.153). The aim is to:
identify and describe conceptions of reality as faithfully as possible /.../ (T)he more faithful we can be to conceptions of an aspect of reality /.../ the better we are able to understand learning, teaching and other kinds of human action.
To us this kind of statement clearly over-reaches any genuine evidence from phenomenographic work. It is an ideologically founded statement, particularly as as yet the ways in which phenomenographic data are produced and analysed is an area of phenomenography which lacks detailed, critical analysis. The injunction of data production processes, understandings and misunderstandings between researchers and subjects concerning the purpose of the research and what its findings represent and can be used and useful for, make such an analysis immediately necessary if we are to judge and put phenomenography to use appropriately.
The characteristics of phenomenographic analysis
As argued, most often the purpose of phenomenographic analysis is pointed out to be the development of categories of description for the different ways in which a thing has been shown to be understood, with these then being accepted as denoting a map of the "collective mind" (Marton, 1995) or the range of possible variations in person-world (or mind-world) relations. However, there is an ongoing debate about what categories of description really are and actually represent. This is not an uncommon situation in relation to the "findings" of qualitative research in general (cf, Denzin & Lincoln, 1994; Beach, 1995). When attempting to define categories of description, using Marton (1995, 175 f.) seems safest:
(W)e are aware of (things) /.../ at the same time albeit not in the same way. Awareness is layered. Some things make up the core /.../ other things belong to the field surrounding the core ...Although we are not consciously aware of most things, we are aware enough for them to be pulled into the core if the changing here and now were to make them appear highly relevant. /.../ Relating parts to parts, parts to whole and whole to context means having a simultaneous focal awareness of parts, whole and context /.../ the major components of the structural aspects of ways of experiencing something /.../ or of conceptions/.../. The structural aspect pre-supposes the referential aspect /.../ of experiencing something /.../ (and) together /.../ characterise /.../ a level of capability (to experience something in a certain way). /.../ (This) way of describing a way of experiencing something is what we call a category of description.
This definition says little about how categories are formed. On the basis of our investigation this involves a repeated reading of transcripts produced for the purpose of phenomenographic analysis after interviews by a researcher. On the basis of these readings certain "utterances" which have been made by research subjects are selected to be grouped together on the basis of what they are said to depict about the phenomenon in question by the reader: basically, what they say to him or her about the phenomenon. However, this is done with little consideration paid to the injunctions of the ways in which the utterances themselves have been formed (cf. Säljö, 1994).
What we have then is "piles" of data, each of which is said to comprise elements (utterances) which, in one way or another, depict something unique about a phenomenon. These piles are examined in terms of their internal consistency (do the elements show consistency in the referential and structural aspect), and the relations between them (do they together provide full coverage of variations in the total data).
That is, the analytic process of phenomenography is an active one, in which data constructions (categories) are formed by the researcher, inspected and used to explore the data in terms of their ability to explain its full range of variations. Two ground rules only are applied: those of internal consistency and parsimony. The minimum number of categories which explains all the variations in the data is the category system taken (parsimony), provided the categories are internally consistent. The ways data are actually shaped in the interactive moment of production are not seriously examined.
We see this as problematic because the meanings and interpretations attributed to subjects' statements can only be designated as meaning denotative or meaning bearing if their meaning elements have been genuinely identified against the background of the shifting totality of meaning comprised within the study. Individual statements should only obtain their phenomenographic meaning in relation to a developing study, not preconceived ideas. We are not sure that this is the case. It seems more to us that meanings are given by the researcher, his/her history/biography and the manipulative situation of the interview conversation.
Phenomenography should be a process of analytic juxtaposition. In this statements should be examined from selected perspectives (which are also accounted for) by the researcher and compared to other statements also examined in these ways. Phenomenographers should look at different statements from the same perspective and the same statement from different perspectives, juxtaposing the outcomes, bringing first one aspect of variation into focus and then another. The aim should be not only to arrive at categories that meet criteria of internal consistency and parsimony and are described by using exemples from the data to substantiate, but also to explore the extent to which the statements which are categorised have been subject to production relations which may potentially at least transform their original subjective meaning they might have carried as well (cf. Larsson, 1986; Francis, 1993).
When dealing with a material in these ways, one will be confronted regularly by variations in the material in the sense that the meanings of statements, expressions, etc., will constantly conflict with each-other, and with researcher understandings of the phenomenon in question, which is also in constant flux. Whilst maintaining some kind of necessarily unambiguous relation to the whole, accomplishing the analysis "correctly" rests on one accepting the fact that variations in meaning and interpretation can occur and be logically inconsistent with each other whilst still being capable of expressing something logical about the phenomenon in question (cf. Smedslund, 1970).
This is recognised by Marton and Booth (1995):
(U)nderstanding and logic are circularly related /.../ what people say is logical given their particular way of seeing the world. /.../ Take /.../ Jenny who claimed /.../ she had 5 fingers on one hand and 10 on the other /.../ (I)t would be very easy to /.../ dismiss her as untaught /.../ (her) expression has the potential to give enormous insight into children's experience of number /.../ When that child's apparently illogical /.../ statement is juxtaposed with other similar /.../ statements /.../ alternative interpretations start to clarify. In the case of Jenny /.../ the fingers /.../ on one hand were called one, two, three, four, five /.../ and /.../ the second six, seven, eight, nine, ten. /.../ (S)he /.../ understands the last number she utters /.../ to be the number of fingers on that hand /.../ Jenny sees number on a counting line of some sort. (Marton & Booth, 1995, p.136)
The implications of the above are that apparent logical and illogical ways of understanding or denoting something can be located through the reading of transcripts in the process of phenomenographic analysis, with the aim then to search the materials for other statements or expressions which might denote the same or similar logicality or illogicality and literally to try to make them make collective sense.
However, this recognition has not been systematised into a recognised form of practice. Phenomenographers categorise in practice only on the basis of what they feel something someone has said about something says about the thing in question. This means that the outcome possibilities of the work are seriously restricted by the perspectives applied in the analysis by the researcher, which, on our reading, are limited to individualistic, cognitivist informed notions of social and cultural mediation processes linking phenomenographic data production and analysis dialectically with the original research assumptions and primary focus. As we also go on to argue, although from a different foundation, this means that there is no genuine consensus method of phenomenography, only a reification.
Phenomenography is research which is said to be concerned with how things are understood, to some degree the experiences of the process of formation of understandings at individual levels, and to some extent their distributions in specific collectivities (also Marton, 1995). In some early research even possible empirical tendencies towards relations between types of learning approach and types of understanding generated were also given a rational basis. Clear methodological guidelines for it are lacking however, as is a clarification of the epistemological and ontological foundations and their consequences. At the end of the day phenomenography appears to be a form of discourse analysis in which the relations of production of discourse and meaning designation are often simply ignored.
Variations: Modes of doing phenomenography
What differs most in phenomenographic investigations as we have read them is what data are produced about, where the point of departure is more or less the same: that concepts can in one way or another be located in distinct human discourse in the terms of which they can also be made sense of. The context of production also varies though, in that the concepts are voiced in different discursive contexts (cf. Säljö, 1994). This has consequences for what is actually "produced" (Säljö, 1994; also Bruner, 1990). In Gothenburg phenomenography, we tentatively argue (cf. Constas, 1992), that there are five recognizable contexts under which the analyzed discourses have been generated and documented. Namely:
- Discursive
- Experimental
- Naturalistic
- Hermeneutic
- Phenomenological
Experimental Phenomenography
In the beginning of the 70:ies "qualitative differences" was the most fundamental concept among the Gothenburg phenomenographers, and often appeared in titles of reports and articles. Here, often under rigorous manipulated conditions, researchers asked students to learn texts from ordinary textbooks or texts of corresponding quality. In subsequent follow ups, which could be repeated several times at various intervals, the focus was primarily directed toward the way students had understood the main point of the text. i.e. They were asked to bend their consciousness inwards (in some way) and to disclose their understanding of what they had read and how they had interpreted, experienced and gone about the learning activity and (implicitly) the development of their understanding. Thus phenomenography started as an experimental enterprise, with interest for outcomes of learning and how students approached the learning task, even conventional measures (e.g. quantitative) of retention, were also used in these experiments.
What was thought of as fundamentally new in this way of investigating learning was the analysis and categorisation of the learning outcome. This usually showed that the texts produced by the readers to account for the original text and its meaning (the learning outcome) were possible to group in a limited number of categories, generally about five or six. This research activity, more than any other, is the one which in one form or another has become associated with phenomenography.
The above procedure is characterised in a classic example of Experimental Phenomenography (Marton, 1975) in which 30 students were asked to read an article of about 1400 words discussing a contemporary reform for Swedish universities. Referring to the content of the article and the message communicated in student responses, Marton (1975, p.275) found the following four ways of understanding the phenomenon 'selective measures', which was the fundamental idea supported by the arguments of the article according to the researcher:
Figure 1. A graphical representation of the hierarchy of levels of outcome (Marton 1975 p.275)
From a phenomenographic viewpoint, the most interesting findings are the qualitatively different ways of understanding the same phenomenon. These constitute the 'outcome space' of the different understandings of a phenomenon. However, as indicated earlier some other parties found how students approached the learning task more interesting. Two fundamentally different approaches were found, surface level processing and deep level processing (Marton, 1975). The difference here concerns whether the student tried to learn the text more or less by heart or if the more fundamental ideas the text gave expression to formed the student's focus of attention.
Other examples of studies in experimental phenomenography include one by Lars Owe Dahlgren (1975) on learning outcomes in economics education, Roger Säljö's study from the same year on "the learner's conception of the learning task" (Säljö, 1975) and Claes-Göran Wenestams investigation of retention (Wenestam, 1980) in which he used 231 subjects.
Discursive Phenomenography
"Why does a bun costs two shillings?" formed the core of an investigation by Lars Owe Dahlgren with a number of school pupils in which he used semi-structured interviews to learn about their conceptions of the phenomenon of price formation (Dahlgren, 1979). Through this work Dahlgren became the first representative for what was later to become called "pure phenomenography" (cf. Marton, 1988; Säljö, 1994), to denote a form of phenomenography which was not directly related (experimentally or otherwise) to an evaluation of the outcomes of specifically pre-directed learning. Dahlgren's interest in price formation had an educational origin, but was treated in the investigation in a way that did not go beyond the knowledge interest of phenomenographic investigation itself. Similar such works are those concerning conceptions of for example social welfare (Wenestam, 1979), nuclear power (Dahlgren & Wenestam, 1980) and death (Wenestam, 1984).
We have chosen the label Discursive Phenomenography rather than pure phenomenography for a couple of different reasons. Firstly of course because we argue that there is no such thing as pure phenomenography, but that rather phenomenography is actually a collection of variegated and pragmatic responses to the demands of investigating a particular kind of research object under different conditions. The label Discursive Phenomenography can be related to this in a number of ways. Firstly to the attributed status of conceptions by phenomenographers, secondly to their ideas concerning the genuine location or residence of conceptions and thirdly in terms of the notion that they can be voiced in a general and context free discourse to be understood non-hermeneutically (cf. Säljö, 1994). That is, Discursive Phenomenography is a direct corollary of phenomenographic assumptions, particularly those of "mid-phase phenomenography" and its concern with mapping conceptions of the world in general of people in general: i.e. it is born of phenomenographic discourse. However, it also uses discourse without regard for the rules of discourse production and analysis to "produce expressions of conceptions" which can be analysed phenomenographically. Discursive Phenomenography might be viewed as the least sophisticated way of doing phenomenography. It can be codified as built upon:
Better known examples of what we term Discursive Phenomenography are Roger Säljö's study on conceptions of learning (Säljö, 1979, 1982), recently followed up by a study by Ference Marton, Elisabeth Beaty and Gloria Dall'Alba (1993). Other examples include studies by Amedeo Giorgi (1986) and Staffan Larsson's investigation of teachers' conceptions of various aspects of their profession and professional role (see Larsson, 1982).
The discursive category has contributed some work which may to a certain extent have miss-credited phenomenography (cf. Säljö, 1994). The idea of being able to use an interview situation to produce a value free discourse in which the genuine understandings of at times quite complex phenomena are voiced by interviewees is problematic. This becomes more so when the researcher lacks in depth knowledge of subject matter related to the phenomenon and of discourse production and analysis. These questions thave not been taken very seriously by the phenomenographers themselves, but may greatly affect the validity of their research disclosures (cf. Kvale, 1989).
Naturalistic Phenomenography
The term naturalistic in Naturalistic Phenomenography alludes to the possibilities for collecting empirical material for phenomenographic analysis from "authentic" situations (cf. Blumer, 1969; Lincoln & Guba, 1985; Berglund, 1983). It is not about producing data in experiments or "extracting" data from speech events specifically arranged for phenomenographic analysis, but about recording what is actually said or happens in a given situation without direct manipulation or involvement from the researcher and then analyzing that data phenomenographically. That is, the idea of this mode of phenomenography is that what is registered as data is also possible to observe "naturally", occurs as part of the routine interactions in a specific (kind of) setting and can therefore be generally representative for other similar such settings.
A good example of Naturalistic Phenomenography is Lybeck's study Archimedes in the classroom (Lybeck, 1981). After experiences in pilot-studies in which he talked to some students, Lybeck entered the life in classrooms to observe and record what happened when instructions in physics dealt with the phenomenon of 'density'. Through recording and analyzing dialogue between students and students and teacher and students respectively, Lybeck established the existence of a number of conceptions of the phenomenon. In the report from the study Lybeck accounts rigorously for his observations as well as for his reflections and analyses during the work. He even accounts for how the different conceptions successively grew more distinct in relation to his deepened understanding of the setting in which they occurred.
The way Lybeck approached the phenomenographic study of conceptions in a school context is one example of how a naturalistic phenomenographic study could be carried out. Whilst a tendency toward Discursive Phenomenography also characterises Lybeck's study, because of the order in which it was done, the naturalistic element may to some extent validate earlier observations.
Hermeneutic Phenomenography
Husserlian phenomenology can be distinguished in terms of three developmental phases, a pre-phenomenological phase which constituted a descriptive psychology of experience, a phenomenological phase embodying the development of a descriptive phenomenology, and a transcendental phenomenology phase representing a contentious stance in phenomenology which gave rise to a bifurcation within it.
The first phase attempted to secure a descriptive psychology as the foundation for all human knowledge, and was thus psychologically reductionist. Mental representations of experiences became the foundations of knowledge rather than material foundations, or a "rain" of ideas. In Brante (1980) this form of psychologism is criticised by exemplification in relation to the construction of understandings of historical events. E.g. explanations of why William the Conqueror didn't invade Scotland, which obviously cannot be limited to mental phenomena alone.
The second phase can be understood by recourse to Husserl's self-criticism in relation to the above. The generalisation of events, experiences and phenomena to single categories strides against their understanding on their own terms. This gave the substance of the second phase of "true phenomenology": understanding things on their own given experiential terms, not in terms of some pre-given reducing general category or categorisation.
The third and perhaps most contentious phase is the transcendental. In a sense developments in this phase represent a turn about by Husserl in relation to his previous critique of Kant and the notion of transcendental ego. Husserl was critical of Kant's infusion of the idea of transcendental ego into philosophy, as an ego which exists outside of experience and links experiences past with those present and future. Husserl's argument was that Kant's philosophy did not require such an ego so there were no reasonable philosophical grounds for it in his treatise. However, he argued, his own phenomenology did need such an ego, as "his" view of consciousness was directional and historical; not just historical and situated, and therefore needed an "organising structure" to direct consciousness itself: something to "bind" experiences and provide a substantive foundation for the intentionality of the mind, for "retention" and "protention".
The introduction of the transcendental ego was however highly contentious in relation to previous phenomenological ontology: which was pointedly non dualist. What happened was that Husserl introduced an element into his philosophy which was outside of the world of experience itself (i.e. outside the temporal flow of experiences of the subject in the world about which we can gain knowledge), and about which direct knowledge cannot be won through experience alone. In this sense it is thus paradoxical, as we cannot win knowledge about it by phenomenological reduction.
This is also where the problematic concept of epoché comes in (see e.g. Beach 1995), as a strategy directing consciousness in such a way to enable knowledge of the transcendental ego to be won. In simple terms epoché is based on the premise that as the world itself is so obvious and cannot be existentially contended, the pursuit of true knowledge is only confounded by our attempts to explain its existence. Epoché represents the reduction of consciousness toward experiences alone and the bracketing of our understandings of the existence of the world: i.e. not making comments about and commitments toward what something is (some experience for instance) until one has adequate experiential grounds for doing so. In short, paying only attention to the content of experiences (noema) and not being concerned with what they represent in existential terms. With respect to providing knowledge about the transcendental ego, bracketing experiences and only directing awareness toward what shows itself in situated experience allows statements to be made about what is also given with the experience but is yet immediately outside it.
The arguments against transcendental phenomenology within the phenomenology movement were strong. Many felt that transcendental phenomenology broke with basic phenomenological tenets by placing an essential philosophical element; the transcendental ego; outside of the stream of experiences making up a subject's experiences of the world. Even Heidegger, Husserl's protegé, argued that phenomenology is essentially a philosophy of existence and that the transcendental ego calls forth the need of epoché which brackets the livedness of existence. But existence, Heidegger claimed, was an essential part of lived experience, beings are in the flow of experiences and are part of the life-world. True phenomenology is an existential philosophy of experience.
This differentiates existential and transcendental phenomenology. In transcendental phenomenology "essences" (the object) are placed analytically outside of phenomenological inquiry. This is accomplished through the bracketting of the experience of being in the world. As the existential phenomenologists see it, the assumption of essence assumes also the situation of the being in the world, so, by bracketing existence from experience a false dichotomy which prevents true phenomenology is introduced (Bengtsson, 1993). The life-world is the fundamental foundation of all knowledge. "Knowing subjects" are of the world and in it, and so too are their experiences.
Modern existential hermeneutics falls inside this kind of tradition. It too operates from the inter-relatedness of being and objectivity (subject-object-subject) relations: although this might not apply to prior forms (cf. Dilthey's and Schliermacher's hermeneutics). Existential hermeneutics asserts that things are never just given of themselves, their meaning must necessarily always be arrived at through interpretation.
Compliance toward the things themselves is an essential aspect of the interpretive act of hermeneutics. Meaning is mediated by being and the formation of lived experience. Both the interpreter and the object of interpretation are part of the same hermeneutic circle of understanding. This is fundamental. The pre-supposition is that distinctions exist between what is given, for instance as an intention in a text written by someone, and its interpretation: the meaning derivation and explification by those who interpret it. Hermeneutics sets out to fuse the horizons of these two meaning formations by bridging the gap between them.
To be true to the object of interpretation, meaning has to be formed as far as is humanly possible in a way true to the original meaning or intentionality of the object: however difficult this might be in practice. This reminds us of comments by Clifford Geertz concerning the aims of such an impossible task as making the pristine interpretation. We try for it even though we know it is not absolutely possible and absolute meaning is never given in any textual formation (cf. also Barthes, 1977). Neverttheless, the basic tenet of hermeneutics is to understand things in their own context and on their own terms, however difficult it may appear.
Whilst we do not feel that this basic tenet is held up in most examples of phenomenography, it is possible to conceive of and there are a couple of possible contenders. In these instances analysis is geared to exegesis, the interpretation of texts or statements not originally made for the purpose of phenomenographic analysis in terms of their whole-part relations. We have termed this phenomenographic form hermeneutic for this reason.
This fundamentally distinguishes this mode from three of the other four categories. An example is a study carried out by Lindblad (1995) which looks at descriptions of pedagogical experiments made by teachers for a School Commission of inquiry in the late 1940:ies. The Commission's aim was to get ideas for changes to curriculum working methods and content as part of the great school reform in the middle of the century. Lindblad does not account extensively for his interpretive work, tending to rather just follow the elaborated routines described by Ödman (1979). He found a number of categories of description for the directions of change indicated by the teachers.
The hermeneutic mode of phenomenography responds better than other forms to the 'humanistic' qualities of a phenomenological perspective. In this sense we feel that this form could benefit from the articulation and clarification of what exactly its relationship to phenomenology and hermeneutics could/should be (see also Uljens, 1993; Marton, 1995).
Phenomenological Phenomenography
Some phenomenographers have striven to identify the phenomenological quality of their work. This can perhaps be most compellingly identified in work by Jan Theman in his study of conceptions of political power (Theman, 1983), Tomas Kroksmark in his interest in didactics (Kroksmark, 1987), Mikael Alaxandersson in his investigation of teacher intentionality and by Michael Uljens in an analysis of phenomenography within a phenomenological framework (Uljens, 1993). These studies are however not typical for ordinary phenomenographic studies. Kroksmark's and Uljens' studies are theoretical, whilst Theman's and Alexanderssons are phenomenological in their approaches the task of finding conceptions.
Dagmar Neuman's study of how young children acquire elementary mathematical skills (Neuman, 1987) is a more typically phenomenographic way of responding to phenomenological criteria. Her studies differ mainly from earlier phenomenographic studies in her elicitation and examination of children's descriptions of experiences of number and the identification of conceptions on the basis of her analyses of these descriptions. Contrary to most other phenomenographic studies in which people are talked with about different phenomena "in general", Neuman asks for descriptions of what is actually going on in the subjects mind during the interview.
This can be compared to Amedeo Giorgi's work in empirical phenomenology (e.g. Giorgi, 1975), as in both Giorgi's and Neuman's cases, the descriptions which are analysed are actually of a first order experiential character rather than, as is normal in phenomenography, second order. Even Roger Säljö's study of conceptions of learning (Säljö, 1979) might be looked upon as a form of empirical phenomenology, although the relation between the interviewee's learning and the text about learning the interviewees have read is far from clear.
A paradox in the context of considering relations between phenomenology and phenomenography is that the early experimental Gothenburg studies carried out before the label phenomenography was developed to refer to them and before attempts were made to explicate the relationship between phenomenology and phenomenography, are those containing some more obvious elements of a kind of phenomenological quality. Although most investigators were most interested in describing 'the outcome of learning', some were more interested in the sense of the ways in which learning was directed toward the meaning of the text as such or toward the fundamental idea the text tried to express. While most phenomenographers tend to view the descriptions of outcomes of learning as the phenomenographic enterprise, phenomenological criteria concern questions directed toward the essences of experiences, such as for instance experiences of learning.
CONCLUSIONS
Phenomenography can be developed in different ways and can embrace thereafter different features and, as we have read and analysed it, has in practice had at least five possible modes of application at Gothenburg. These are: (i) the discursive, (ii) the experimental, (iii) the naturalistic, (iv) the hermeneutic, and (v) the phenomenological.
The identification of the five modes of phenomenographic research took its starting point in a critical reading of actual phenomenographic works. But it didn't altogether stop there. It also displayed the pragmatic foundations of the approaches to data-generation, and their consequences. The five modes of phenomenography together offer an alternative understanding of what phenomenography actually is about. Indeed, the variations indicate that it may be about different things, and that this may be hidden from view behind a label and because of reification and a lack of methodological reflexivity amongst phenomenographers. Two of the modes have been indicated to have some developmental possibilities in relation to existential phenomenology and hermeneutics.
"Good-for-nothing"? Needleman (1963) judged phenomenography as "good-for-nothing", yet phenomenography in the ways it has been practiced at Gothenburg has created a long and productive research tradition. A lot of empirical data have been generated and in some cases these have been put to practical use (cf. Pramling, 1994; Bowden, 1995). However, the argument is that most phenomenography is not used (Hasselgren, 1993). A contribution to this problem is the uncertainty of what the bases of phenomenography as practiced actually are. Marton wants to see phenomenography as empirical phenomenology (Marton & Neuman, 1990). Our argument is that this is relevant to very few examples of phenomenography and may even restrict the development of desirable research variation.
Unlike Needleman we are not of the view that phenomenography has to be phenomenological, nor that it is good for nothing. Our investigations show that it has the potential to provide useful information in investigations of learning, learning outcomes and experiences of learning processes etc., providing the injunctions of method are explored critically by practicing phenomenographers in relation to what they do, how they do it and how this bears on research disclosure possibilities. As an ideational platform, phenomenography has started a much needed wave of renewed interest in curriculum research in Sweden (cf. Marton, 1986; Kroksmark & Marton, 1987; Lundgren, 1987; Ahlström & Hasselgren, 1988).
Brother of phenomenology? As is evident from the previous descriptions of the five modes of phenomenography, some of these modes are far from what might be conceived of as phenomenological. For instance, an experimental context Experimental Phenomenography is not one in which you meet your dialogue partner unbiased, nor do you do this when compiling fragments of discourse which are not hermeneutically interpreted Discursive Phenomenography and unbias is a necessary precondition of phenomenology
Similar problems entrap Naturalistic Phenomenography as well. From naturalistic points of departure, the position and perspectives of other participants in the social situations in which the phenomenography has been done are always considered, as is the researchers understanding of the relations of different definitions of situation (e.g. Lybeck, 1981). This is not so evident in naturalistic phenomenography. Two modes remain, the hermeneutic and the phenomenological, and it has been indicated in this paper that in some senses these two modes of phenomenography do evidence some developmental possibilities as forms of phenomenology or may benefit from the articulation of a research base in phenomenological and other forms of existential philosophy.
In short, two of five modes of phenomenography show some phenomenological qualities. Quantitatively a lesser part of phenomenographic research does so in practice. The conclusion is that Gothenburg phenomenography is not good for nothing, nor a brother of phenomenology. It is productive research which can be developed in a number of ways, two of which may in some way be relateable to phenomenology.
NOTE
There are three versions of this text, (i) the original paper written to the NFPF - Nordic Society for Educational Research - Congress in Vasa, Finland 1994 (Hasselgren, 1994), (ii) this version of the Vasa paper, (iii) an article in the special issue on phenomenography in the Australian journal Higher Education and Research Development (Hasselgren & Beach, forthcomming). Our reason for publishing this version is that it contains some intersting material not contained in the two shorter versions. Version (i) has no sub-title, version (ii) has the sub-title Or: Phenomenography is what Phenomenographers do when doing Phenomenography, and version (iii) has the sub-title; Outline of an Analysis, which we think mirrors the development of the papers in an adequate manner.
The research project "Kartläggning av fenoenografin" (Mapping the Phenomenography) - KAFE, financed by the Swedish National Agency of Education, is the framework within wich this paper has been produced.
The "alternative" title of this paper "Phenomenography is what phenomenographers do when doing phenomenography" alludes to the title of the expert opinion given by ke Bjerstedt (1986) concerning Lars Owe Dahlgren's qualifications for a professorship in education/didactics entitled What does a phenomenographer do when doing phenomenography?
LITERATURE
Ahlberg, A. 1992: Att möta matematiska problem. En belysning av barns lärande. Göteborg: Acta Universitatis Gothoburgensis
Ahlström, K-G. & B. Hasselgren, 1988: Didaktikvågen i svensk pedagogik. Nordisk Pedagogik, 8, 1-3.
Alexandersson, M. 1994: Metod och medvetande. Göteborg: Acta Universitatis Gothoburgensis
Barthes, R. 1977: Image, music, text. New York: Hill and Wang
Beach, D. 1995: Making Sense of the Problems of Change: An Ethnographic Study of a Teacher Education Reform. Göteborg: Acta Universitatis Gothoburgensis
Bengtsson, J. 1993: Sammanflätningar. Husserls och Merleaud Pontys fenomenologi. Göteborg: Daidalos
Berglund, G. 1983: Om naturalistisk forskning. Någraanteckningar. Pedagogiska institutionen, Uppsala universitet, Arbetsrapporter nr 68.
Bjerstedt, Å. 1986: Vad gör en fenomenograf när han fenomenograferar? Några glimtar från ett pedagogiskt författarskap. Institutionen för pedagogik, Lärarhögskolan, Malmö, Småtryck och särtryck nr 560.
Blumer, H. 1969: Symbolic interactionism: Perspective and method. Englewood Cliffs: Prentice Hall International
Booth, S. 1992: Learning to program. A phenomenographic perspective. Göteborg: Acta Universitatis Gothoburgensis
Bowden, J. 1995: Phenomenographic research: some methodological issues. Nordisk Pedagogik, 1995/3, 145-155.
Brante, T. 1984: Vetenskapens struktur och förändring. Lund: Doxa
Bruner, J. 1990: Acts of meaning. London: Harvard University Press
Constas, M. 1992: Qualitative analysis as a public event: The documentation of category development procedures. American Educational Research Journal, 29, 253-266.
Dahlgren, L.O. 1975: Qualitative differences in learning as a function of contentoriented guidance. Göteborg: Acta Universitatis Gothoburgensis
Dahlgren, L-O. 1979: Children's conception of price as a function of questions asked. Institute of Education, University of Göteborg, Report No. 81.
Dahlgren, L.O. & C-G. Wenestam, 1980: Ideas about nuclear power plants. Department of Education, University of Göteborg.
Denzin, N. & Lincoln, Y. 1994: Entering the field of qualitative research. In Denzin N.E. and Lincoln Y.S. (1994: eds.) Handbook of Qualitative Research. London: Sage
Francis, H. 1993: Advancing phenomenography: Questions of method. Nordisk Pedagogik, 1993/2, 68-75.
Giorgi, A. 1975: An application of phenomenological method in psychology. I Giorgi, Fisher & Murray (red): Duquesne studies in phenomenological psychology. Pittsburgh: Duquesne University Press
Giorgi, A. 1986: A phenomenological analysis of descriptions of learning obtained from a phenomenographic perspectiv. Institutionen för pedagogik, Göteborgs universitet, publikation nr 1986:6.
Hasselgren, B. 1993: Kartläggning av fenomenografin (KAFE) förslag till forskningsprojekt. Institutionen för pedagogik, Göteborgs universitet.
Hasselgren, B. 1994: Phenomenography A ²good-for-nothing brother² of phenomenology? Paper presented at the NFPF Conference in Vasa, Finland, 1994.03.10-13.
Hasselgren, B. & D. Beach, (forthcomming): Phenomenography Ð A Ógood-for-nothing brotherÓ of phenomenology?Outline of an analysis. Journal of Higher Educationa and Research Development.
Johansson, B., F. Marton & L. Svensson, 1985: An approach to describing learning as a change between qualitatively different conceptions. I Pines & West (red): Cognitive structure and conceptual change. New York: Academic Press
Kroksmark, T. 1987: Fenomenografisk didaktik. Göteborg: Acta Universitatis Gothoburgensis
Kroksmark, T. & F. Marton, 1987: Läran om undervisning. Forskning om utbildning, 3/1987.
Kvale, S. 1989: To validate is to question. In S. Kvale (ed.): Issues of validity in qualitative research. Lund: Studentlitteratur
Larsson, S. 1982. Studier i lärares omvärldsuppfattning: Ett försök till integration. Rapporter från pedagogiska institutionen, Göteborgs universitet, 1982:15.
Larsson, S. 1986. Kvalitativ analys exemplet fenomenografi. Lund: Studentlitteratur
Lincoln, Y. & E. Guba, 1985: Naturalistic inquiry. Beverly Hills: Sage Publications
Lindblad, S. 1995: Lärarna samhället och skolans utveckling. Stockholm: HLS Förlag
Lundgren, U. 1987: Didaktikens namn. Forskning om utbildning, 3/1987.
Lybeck, L. 1981: Arkimedes i klassen. En ämnespedagogisk berättelse. Göteborg: Acta Universitatis Gothoburgensis
Marcus, G. 1994: What comes after ²post²? In Denzin & Guba (Eds.): Handbook of qualitative research. London: Sage
Marton, F. 1975: On non-verbatim learning: Level of processing and level of outcome. Scandinavian Journal of Education, 16, 273-279.
Marton, F. 1981: Phenomenography Describing conceptions of the world around us. Instructional Science, 10, 177-200.
Marton, F. 1986: Vad är fackdidaktik? I F. Marton (red): Fackdidaktik. Vol I. Lund: Studentlitteratur
Marton, F. 1988: Phenomenography: Exploring different conceptions of reality. In D. Fetterman (ed.): Qualitative approaches to evaluation in education. The silent scientific revolution. New York: Praeger
Marton, F. 1992: Phenomenography and "the art of teaching all things to all men". Journal of Qualitative Studies in Education, 5, 253-267.
Marton, F. 1995: Cognosco ergo sum. Reflections on reflections. Nordisk Pedagogik, 1995/3, 165-180.
Marton, F. & S. Booth, 1995: Changing approaches to learning. Department of Education and Educational Research, Göteborg University. (manuscript)
Marton, F. & D. Neuman, 1990: Constructivism, phenomenology, and the origin of arethmetic skills. I Steffe & Woods (red): Transforming childrens mathematics education. International perspectives. Hove: Lawrence Earlbaum Associates
Marton, F., G. Dall'Alba & E. Beaty, 1993: Conceptions of learning. Journal of Educational Research, 19, 277-299.
Marton, F., L.O. Dahlgren, L. Svensson & R. Säljö, 1977: Inlärning och omvärldsuppfattning. Stockholm: AWE Gebers Nationalencyklopedin, 1991, vol. 6. Höganäs: Bokförlaget Bra Böcker
Needleman, J. 1963: Beeing-in-the-world. Selected pepers of Ludwig Binswanger. New York: Souvenir press
Neuman, D. 1987: The origin of arethmetic skills. A phenomenographic approach. Göteborg: Acta Universitatis Gothoburgensis Pramling, I. 1994: Fenomenografi och pedagogisk praktik. Nordisk Pedagogik, 1994/4, 227-239.
Rainbow, P. 1986: Representations are social facts. Modernity and post-modernity in anthropology. In Clifford & Marcus (Eds.): Writing culture: The poetics and politics ethnography. Berkley: University of California Press
Sandberg, J. 1995: Are phenomenological results reliable? Nordisk Pedagogik, 1995/3, 157-164.
Smedslund, J. 1970: Circular relation between understanding. Scandinavian Journal of Psychology, 11, 211-219. Sonnemann, U. 1954: Existence and therapy. An introduction to phenomenological psychology and existential analysis. New York: Grune & Stratton
Säljö, R. 1975: Qualitative differences in learning as a function of the learner's conception of the task. Göteborg: Acta Universitatis Gothoburgensis
Säljö, R. 1979: Learning in the learner's perspective. I: Some common-sense conceptions. Report from the Department of Education, University of Göteborg, No. 76.
Säljö, R. 1982: Learning and understanding. A study of differences in constructing meaning from a text. Göteborg: Actas Universitatis Gothoburgensis
Säljö, R. 1994: Minding action: concieving of the world versus participating in cultural practices. Nordisk Pedagogik, 1994/2, 71- 80.
Theman, J. 1983: Uppfattningar av politiska makt. Göteborg: Acta Universitatis Gothoburgensis
Uljens, M. 1989: Fenomenografi forskning om uppfattningar. Lund: Studentlitteratur
Uljens, M. 1993: The essence and existence of phenomenography. Nordisk Pedagogik, 13, 134-147.
Wenestam, C-G. 1979: Social vård i en social värld. Hur mannen på gatan uppfattar vissa aspekter av socialvården. Pedagogiska institutionen, Göteborgs universitet, rapport nr 188.
Wenestam, C-G. 1980: Qualitative differences in retention. Göteborg: Acta Universitatis Gothoburgensis
Wenestam, C-G. 1984: Qualitative age-related differences in the meaning of the word "death". Death Education, 8, 333-347.
Wolcott, H. 1992: Posturing in qualitative inquiry. In leCompte, M. D. Millroy, W. L. and Preissle, J. (1992: Eds.). The Handbook of Qualitative Research in Education. New York: Academic Press
Ödman, P-J. 1979: Tolkning, förståelse, vetande. Hermeneutik i teori och praktik. Stockholm: AWE/Gebers