Berufliche Bildung in Europa – 20 Jahre nach Lissabon und am Ende von ET 2020. Entwicklungen und Herausforderungen zwischen supranationalen Strategien und nationalen Traditionen

1 Background 

In 2016, the Austrian National Qualifications Framework (NQF) Act came into force. Its development was inspired by the EU Recommendation for a European Qualifications Framework (EQF). As the EQF is a recommendation, there is a certain amount of freedom in national implementation, on the one hand whether the recommendation is implemented at all, and if so, whether this is done in full or in part. Community law also provides scope for implementation to take account of the respective legal competences and circumstances as well as the traditions of the member state. However, from the perspective of EU law, the member states are not completely free to design a national qualifications framework. And under no circumstances should a decision to implement it frustrate or jeopardise the harmonisation purpose of the Recommendation (Herdegen 2011, 228).

The structural features of the Austrian Qualifications Framework correspond in the central design elements with those from the EU Recommendation. For example, the basic idea of the EQF, namely to use learning outcome as a decisive characteristic for level assignment, was adopted, although for Austria this requires a new, little practised perspective on the comparison of qualifications. The Austrian Qualifications Framework also has eight levels and the German-language descriptors of the levels in the wording of the EQF Recommendation (from 2006) were adopted in national law (an exception are the academic degrees awarded by higher education institutions). The levels are described accordingly in the differentiation of knowledge-skills-competence. Furthermore, the aim is to represent the entire learning continuum (formal, non-formal and informal learning), thus potentially opening the NQF to all segments of the Austrian qualifications landscape. The referencing to the EQF is a formality with the design of eight levels and the largely recommendation-based design of the NQF. At the same time, this procedure for referencing a set of levels in a national qualifications system to those in the EQF does not force to lay down criteria for the case of imperfect fit. This will have to be clarified later.

Overall, however, this is still a highly ambitious programme, considering that Austria's education system is characterised by very fragmented responsibilities and different internal logics. These range from a very traditional company-based model of apprenticeship training, which is largely implemented in a work-integrated manner, a broadly developed vocational school system characterised by practical, theoretical and general education teaching, the higher education system, which pursues independent, research-based curricular principles in relative autonomy, the health system, in whose training regulations an expert model is inscribed, and the field of CET, which operates largely detached from legal requirements.

For the allocation of qualifications to levels, the law basically defines two approaches. The academic degrees (according to the Bologna study architecture) for which the Dublin descriptors are relevant have been assigned directly to levels six to eight by law (§3,2). Thus, there is a specific alignment of levels in the NQF with those of the framework of qualifications for the European Higher Education Area (QF-EHEA). For all other qualifications, the law stipulates that individual assignment procedures must be completed. A distinction is made here as to whether the qualification is subject to statutory regulation, i.e. a national or regional authority has the professional standards or curricula, or not. In the first case, the application for assignment must be submitted by the competent authority. For the second case, the NQF Act stipulates (§9) that this must be done by specially qualified bodies authorised by the Ministry of Education (NQF service bodies). Following an expression of interest and application, six such bodies were designated in May 2019 and they have already started their work - with varying degrees of intensity.

The assignment procedures to the NQF were started in the formal sector from 2017 and qualifications from the non-formal sector were assigned for the first time in June 2020. The assignment requests are complex application procedures which, in addition to descriptions of the qualification and the submission of corresponding evidence, must be based centrally:

• To provide a description of key learning outcomes of subject-specific and generic learning outcomes related to the qualification.
• Disclose the mode of assessment or examination procedure and criteria of the assessment procedure.
• To present a justification of the proposal for assignment to one of the NQF levels against the background of the level-specific descriptors.

The assignment requests are formally and technically examined and finally submitted to the national steering group for a final decision. Representatives of the qualifications landscape are broadly represented in the steering group. All federal ministries, the Länder, social partner organisations and representatives of selected education sectors are delegated to this body. If this steering group does not prohibit an assignment to the desired level by a qualified majority (two-thirds) (veto right), the entry of the qualification in the publicly accessible qualification register is made at the level for which the argument was made.

In the course of the formal and substantive examination, a wide range of expertise is drawn upon. The procedural expertise lies with a National Coordination Point (NCP), which is the NQF's administrative office. The content expertise is essentially provided by a body of experts, the NQF Advisory Board, which comments in writing on each assessment. This advisory board consists of seven experts (regulated by law), including at least one expert from the field of health care. The members of the advisory board must have outstanding professional qualifications in the fields of professional practice as well as education, further education and training. They are appointed by the Federal Minister of Education, Science and Research. This appointment is made considering proposals from the NCP, the Advisory Council for Economic and Social Affairs (the central board of social partner organisations) and the Agency for Quality Assurance and Accreditation Austria (national quality agency for higher education) of two experts each, as well as from the Federal Ministry of Health of one expert. These proposals require the approval of the NQF steering group by a simple majority vote before appointment. For individual assignment procedures, expert opinions may also be obtained. In doing so, the NCP can rely on a list of experts agreed by the steering group.

This complex organisational structure became necessary because:

• by definition of the law, there is no competence in the public administration for non-formal qualifications and thus an allocation procedure only could be based solely on bureaucratic action (clear responsibilities) they would be excluded.
• overarching criteria for the different sectoral principles and traditions do not seem possible
• thus, good conditions for factually justified allocations exist and a public arena for balancing different professional, sectoral and political interests is given.

Thus, different negotiation arenas were created for political and factual aspects, which is due to the claim that allocation procedures should not be left exclusively to procedural pragmatism or a power play of political actors (Spöttl 2012, 225). This model of steering was taken from the scientific monitoring of the implementation process and with certain adjustments (tendency to weaken the expertise-oriented advisory board) was adopted in the legal norm. The development work was also flanked by a gratifyingly broad consultation process. While in the course of the development work there were undeniably important aspects discussed, such as the diagnostic reliability of assessment procedures, now, after the pioneering phase, other, but much more fundamental questions emerge. A closer look reveals challenges precisely within this expertise-based model of assignment that are worthy of further investigation.

2 Fundamental questions remain

Two – non trivial – questions have remained unanswered in Austria to date lead to uncertainties in the concrete assignment procedures, a fuzziness of the justifications for (non-)assignments, to uncertainties among the applicants and subsequently to questions of the sustainable acceptance of the NQF in the (expert) public. These two central and interrelated aspects of the assignment work are, it must be noted, already to be seen as unclear or underdetermined in the recommendation on the EQF and were thus imported into the national implementation, without, however, having been formulated more explicitly there:

1. What rationale underlies an assignment to a level? Can classification in the narrower sense really be assumed, even if this would be the exact designation in the EQF recommendation?
2. What exactly is meant by learning outcome (LO) as the object of classification?

This may come as a surprise, since both terms are core elements of the EQF as well as the NQF. But apparently the seemingly intuitive words learning outcome and classification, as well as the rather formalistic definitions of the EU Recommendation on this, obscure the deep structure of the problem. From a political realism point of view, it must be said that this vagueness or ambiguity will to some extents have been a prerequisite for acceptance among the member states.

2.1 Classification as a non-trivial procedure

The stipulation that NQFs are per se classification instruments ("national qualifications framework' means an instrument for the classification of qualifications", EU 2017, Annex I), which is put into function "according to a set of criteria for specified levels of learning", raises (at least) two questions. The first is the admissibility of the assumption of the social or intersubjective robustness of units that are categorised. In this context, cognitive linguistics has long problematised a "folk theory of categorisation" which "says that things come in well-defined kinds, that the kinds are characterised by shared properties, and that there is one right taxonomy of the kinds" (Lakoff 1987, 121). However, we will not pursue this further here.

Secondly, and this should be of special interest here, the principles according to which membership to one of the classes is argued need to be clarified. And indeed, in various scientific disciplines, which, like modern biology, are based entirely on corresponding taxonomic models as a framework theory, corresponding questions have been negotiated in a differentiated manner. Overall, it can be shown that there is no uniform model of classification in the scholarly discussion. During the history of philosophy and science several definitions and concepts of classification have been put forward. A surprisingly long chronological list of definitions appears in Birger (2017). Thus, Lakoff's assessment that, beginning with Aristotle, the classical theory of categories (and classifications) was regarded as unproblematic (Lakoff 1987, 6) can only be followed to a limited extent, because his assessment remains strongly focused on attributions of meaning and basic epistemological attitudes and ignores special discussions, for example, in the life sciences (from Linne onwards) or library science. But he is certainly right that in recent decades additional understandings have been added to traditional feature semantics and, in the context of classifications, new understandings that go beyond abstract containers being filled with things of different features clearly according to belonging or non-belonging and that objects and designations could be divided in analytical processes into parts and partial meanings with sharp edges. In this context, prototype theory, Familienähnlichkeit, contextual or instrumental theory of meaning, etc. can be mentioned. Logical problems with feature extraction in the classical way have turned out to be that it is often not possible to generate (closed) systematic catalogues of features and to satisfactorily come to grips with the phenomena of continuous transitions between features (no exclusive binarity).

If we look at the debate in biology, which has been going on for a long time, we see that the concept of classification was used there for a certain form of order creation. However, under the umbrella term of ordering (taxonomy), this has been supplemented by a concept that is now more significant for this scientific discipline, namely that of the system. While a classification is based on purely formal characteristics, or more precisely on similarities of these characteristics, a system, on the other hand, expresses the inner relation of objects.

2.2 Discontinuous variation as a starting point

The starting point for any taxonomy is first of all the observation that the diversity of forms (for biology: the found forms of living beings) did not appear disorderly. At the same time, however, that they do not form a continuous and certainly not a random distribution, but rather "discontinuous variation" (Bateson 1891, 159), but at the same time "a group-wise arrangement of certain trait distributions" (Dobzhansky 1937, 214) is in place. Uniformity within groups and non-uniformity between these groups determined the overall picture. And ultimately, as opposed to a strictly trait-based classification model, the relational element of system formation, the degree of natural connection, (genetic) kinship or descent prevailed as a basic taxonomic principle in a systemic perspective.

Detached from a disciplinary perspective, Peirce names the corresponding effort as an intellectual achievement as early as 1902, years before the theory of evolution had established itself as a metatheory of biology: "No greater merit can a taxonomist have than that of having his eyes open to the ideas in nature; no more deplorable blindness can affect him than that of not seeing that there are ideas in nature which determine the existence of objects" (Peirce 1931, 103). And he speaks out against an atomistic understanding of science as a description of the world as a sum of objects with concretely nameable characteristics: "A science is defined by its problem; and its problem is clearly formulated on the basis of abstract science" (ibid., 102). As a result, differences that refer to individual characteristics can be overcome. And in the formation of order or grouping of elements (taxonomy), classification (ordering of elements based on their intrinsic characteristics, e.g., similarity) is replaced by systematics (ordering of elements based on their relations within a natural process, e.g., descent).

This is also reflected in the context of the EQF Recommendation, which refers to the "integrating and coordinating national qualifications subsystems" by means of national classification. The inner relationship in these subsystems would therefore have to be included in the consideration of LO and to refrain from an atomistic interpretation of descriptions and terms. It must also be borne in mind that the relations in different subsystems can have different characters and justifications and can also differ with regard to continuity, flowing transitions, even progress, breaks and beginnings and ends.

If this circumstance is not taken into account by means of internally differentiating sub-frameworks (with sector-specific logics) within an NQF, the assessment of complexity and interrelationships of learning outcome descriptions presents itself as a challenging task that does not seem to be manageable with uniform, cross-sectoral criteria catalogues. And this already touches on the second question raised, what exactly is designated with LO.

2.3 Theory effects in connection with outcome concepts

The disarmingly simple argument of using the outcome of qualification processes for Europe-wide comparison, or at least preferring it to the hitherto weightier input characteristics or bilateral comparison procedures or credits, is placed in the educational science debate by the chosen terminology of learning outcome and should no longer be understood in everyday language or used naively. This concerns curriculum-practical, curriculum-theoretical but also education-political aspects.

For the latter, reference may be made to Basil Bernstein's contribution, which already dates back some time, in which he shows how, via the classification and framing of knowledge, a society selects, classifies, distributes, transmits and evaluates the educational knowledge, power is exercised in this way and principles of social control become effective. (Bernstein 1971) But this would go too far at this point. In the following, we will deal with the two curriculum-related questions. As will be shown, these cannot be viewed separately, but show interdependencies.

Much has already been said and written about the aspect of LO, be it as a "key concept in a changing education policy landscape" (Prøitz 2015, 275). But it is also noticeable that there is a certain obliviousness to history in the consideration of the concept. There is no other way to understand how a substantial study of scholarly understandings of the LO concept can conclude that "[t]he history of the term "learning outcome" is short" (ibid., 121) and makes do with marginal references (to Robert Gagné and Elliot Eisner in the 1970s respectively) and only a passing reference to the learning objectives taxonomy. And the debate essentially relates to the field of curriculum development, which would have gained momentum beginning in the 1990s. But even in the literature there are enlightening findings such as Adam's that "there are possibly confusions between learning outcomes, objectives and aims" (Adam 2004, 19), which will be discussed in the following.

For the German-speaking world, the already quite developed debate on the differences between learning objectives and (intended) learning outcomes may be recalled at this point. While learning outcomes (at least those that are described at the concrete teaching-learning level) describe observable behaviour, goals can be more comprehensive or integrated and describe changes in behaviour over time. But what again requires operationalisation (and partly hierarchisation) is to check the degree of goal achievement later. If this difference is not sufficiently taken into account, unintended effects can occur. Blankertz recognised this early and identified possible consequences (1975, 161, transl. PS):

• Exaggerating the predictability of learning outcomes.
• Overlooking the logical contradiction between basic pedagogical goals, which intend the learner to be released to himself, and determination through complete operationalisation;
• Levelling the different subject and academic structures through behavioural dispositions and reinvigorating them through the most naive form of so-called "formal education";
• To fortify the wrong question for curriculum development – giving either goals or content the forefront – instead of going back to the criteria question (the question of why and how in operationalisation).

However, these risks are not reasons to exclude operationalisation, but rather regulative principles for quality-oriented curriculum development, which wants to distinguish clearly between goals and results. Curriculum development can be broken down into actual events and described essentially as a process of (structural and content-related) decisions. However, it is important to bear in mind that 'content' or 'educational goal(s)' are not trivial or one-dimensional bodies of knowledge that can be combined at will.

In particular, a clear differentiation between the teaching-learning level and the outcome-oriented description of qualifications seems highly significant. A circumstance that should actually belong to the disciplinary core of the educational science discussion since Ralph W. Tyler's authoritative contribution (Basic Principles of Curriculum and Instruction from the 1949) but is mixed up in the EQF and NQF context with requirement-related qualification perspectives.

While Tyler focused on "defining the objectives", "organising learning experience", "effective instruction" and "determining the defined objectives", the qualification-oriented debate predominantly focuses on an understanding of LO as standards.

3 Are we playing a losing game?

Possibly a transitional problem is being addressed here, since in future all qualifications will already be designed for a certain NQF level and prescriptive curriculum development or revision will take place with regard to certain NQF levels. But even in the case that this will indeed be the case, the duration of curricular processes has to be taken into account and then we are talking about many years of transition.

The evident indeterminacy of classification processes is surprising against the background of the identification of NQFs in the European Qualifications Framework Recommendation. There is no indication what procedures, methods or principles this process can and should be based on. This can perhaps be formally justified with the circumstance that this is not primary law of the European Union and should interfere as little as possible with the sovereignty of the member states. However, this relevant interface for the diagnostic reliability of referencing national qualifications to the EQF, seems to be considerably underdetermined. A general reference to learning outcomes that have to be assessed according to a fit to the descriptors of the levels is not sufficient in its degree of determination. This is especially true since the descriptors - how could it be otherwise if the most diverse sectors are to be integrable - have a generic character.

In the development process and the current accompanying documents, reference was and is made repeatedly to the so-called best-fit principle, which is supposed to be essential (EU 2013, 49). However, the referencing process of the NQFs to the EQF dominates in the current documents. The justification that this would be an accepted procedure in various sciences (mathematics and engineering) is linked to the argument that there can legitimately be differences in categories of level descriptors and numbers of levels in the NQFs. But in a departure from rigid criterion-led principles, reference is made to a plausible but different consensus model that "requires a common judgement from a range of stakeholders so that there can be confidence in the outcome of the approximation (EU 2013, 50).

In the European Recommendation, as well as in the Austrian law, there is no information on a concept of best-fit. In the course of the public consultation in the run-up to the development of the NQF, very critical voices were heard on this from the university sector, which, however, had aimed at a balanced weighting of the Knowledge-Skills-Competence (KSC) dimensions in the assignment of qualifications to the NQF in order to prevent the highest fit of one of these dimensions, detached from the other dimensions. Paradoxically, the universities, or more precisely their graduates, would also be affected if this principle were followed consistently. For compared to the undeniable strengths in the area of knowledge, the other two descriptive dimensions lag behind, in some cases considerably.

If, at the national level, sub-sectors that have so far been linked to different principles and traditions, shall be related to the NQF problem arise, because the sector-specific characteristics are often not made explicit, or their internal classification logic does not seem transferable. However, it must be noted that in Austria – apart from the European Higher Education Framework (EHEF) – there are hardly any explicit sectoral QFs. Certain (explicit or immanent) levels of qualifications already exist in the education system (school system, language reference frameworks, ...), sometimes also regulated by law, but rarely form multi-level models; instead, qualification-specific (often formal) differences are named. Also, learning outcome orientation is (so far) seldom developed to such an extent that derivations can be made here.

Assignments can become problematic if no overarching criteria are (or can be) named and this is done either from power- or dominance-specific positions (see the reference earlier to Bernstein) or these assessments are subordinated to a supposed common-sense. Common sense means here that "[w]e naturally feel that [we all] embody self-evident laws of thought, the same for all men. We know all the answers!" (Whorf 1956, 238). But, Whorf continues, "But, when scrutinized, they become dusty answers." So, one could diagnose that we are playing a losing game? At least as far as we trust in reliable and factually well-founded categorisations.

In terms of absolute results, that will probably be the case. But if common sense is thought of in a somewhat broader form, in the direction of a factual and at the same time multi-perspectival, discursively produced assessment, legitimate results could be achieved. To this end, we must admit that the various sub-sectors of education and training, which differ in traditions, languages, concepts, etc., turn out to be parties in a negotiation process of overarching representation.

In this way, however, previously untapped substantial potential, for example in domain specificity, lifeworld orientation, non-cognitivist learning theories or even vocational pedagogical expertise of trainers and teachers, who could represent assets of vocational education in such procedures of identifying observable behaviour on the one hand and social learning on the other, can be made fruitful for the classification debates.

In this respect, the conception of the transdisciplinary advisory board within the framework of the Austrian procedure, which systematically provides a discursive arena that creates representation for subsystems – albeit limited for pragmatic reasons – can be described as a rational path. At least if louder, agreement-oriented debates are conducted there, which are not counteracted by upstream or downstream decision elements (with other logics) of the allocation procedure. The future will show whether this will succeed.