Liam Magee

S: But you always need to put things into a context, don’t you?

P: I have never understood what context meant, no. A frame makes a picture look nicer, it may direct the gaze better, increase the value, but it doesn’t add anything to the picture. The frame, or the context, is precisely the sum of factors that make no difference to the data, what is common knowledge about it. If I were you, I would abstain from frameworks altogether. Just describe the state of affairs at hand (Latour 2004, p. 64).

What to measure—describing ‘ontological cultures’

Having worked through a range of theoretical positions in the preceding chapter, it is now possible to put forward our own model of the kinds of conceptual entities which are both explicit in knowledge systems themselves, and implicit in the practices and beliefs of the people who design and use them. These ‘entities’ have so far been described through a series of near-cognate, proximally synonymous terms, ushered in throughout this study to denote both a given system or arrangement of concepts, and, on occasion, also the social environment, and the people who produce and consume them. Yet none of these terms—perspectives, world views, paradigms, epistemes, conceptual schemes or spaces, historical ontologies, lifeworld, habitus—seems quite adequate for the kinds of entities wanted here. The following account aims to characterise these entities in descriptive terms, before then offering a formalised account as a part of the framework further on below. The account may seem more dogmatic that is intended; just as Minsky notes in a similar framework endeavour, this account proceeds while ‘pretending to have a unified, coherent theory’ (Minsky 1974).

What is envisioned here, then, is an elastic, dynamic, fluid yet interconnected ‘structure’ shared across members of a group or organisation; neither a subjective, individual cognitive ‘lifeworld’, nor a stable, socialised epistemic ‘system’, but something at an intermediate and intersubjective level of granularity. ‘Conceptual scheme’ seems adequate though insufficient, as the sought-after concept must also embrace the structural conditions and social practices that give rise to such schemes. Stopping at the conceptual misses out on these elements. A more embracing term is needed, which directs attention out from subjective cognitive abstractions towards the objective and intersubjective spheres in which those abstractions are generated, and to which they correspond.

‘Culture’ is one possible term; it both signifies a collective group and, more remotely, connotes a homogenous, self-replicating organism. The term has the advantage of being at the right granular level, since it is elastic, and can be stretched and scaled along several dimensions; it can describe a large or small, short or long-lived, casually or formally, historically or spatially organised collective of individuals. One of the functions of the cultures considered here, though, is that they produce very particular kinds of artefacts—formal knowledge systems. To describe just those cultures engaged in the production of knowledge systems particularly, I have added the epithet ‘ontological’. Taken literally, an ‘ontological culture’, then, is something which produces formal knowledge systems like semantic web ontologies—organisations, communities and other social groups who, as one of their practices, organise slices of the world into classificatory schemes. More tenuously, ‘ontological culture’ can also be taken in several other senses too: a culture which, to coin a neologism, ontologises; actively constituting its world and the beings in it (meaning something similar to Hacking’s use of the phrase ‘historical ontology’); or even as a biological ‘culture’, which is differentiated from more mundane microbial kinds of culture by being ‘ontological’ in the philosophical sense. The conjoined term, as a result, operates as a weak double pun, implying each of these meanings. Though concisely descriptive, this term does however strain at convention use; occasionally through the study the more conventional term ‘knowledge cultures’ has been preferred—though lacking in specificity the latter term also has an existing resonance in the sociology of knowledge and science, for instance, Knorr-Cetina (1999) and Peters and Besley (2006).

An ‘ontological culture’ inherits many of the characteristics ordinarily assigned to cultures generally. The remainder of this section presents a basic narrative, unfolding a series of terms as it develops a description of ‘ontological culture’. These terms, in turn, are formalised into a more coherent model, which is employed in the more technical discussion of the framework further on.

The organic connotation of ‘culture’ implies a certain autonomy—that cultures are, like Luhmannian systems, first-order sociological, rather than psychological or biological entities. They are in some sense irreducible to the agents or actors who comprise them. Actors instead perform semi-deterministic roles in accordance with the functional goals of a culture, of which there can be many: for example, generating profits, delivering services, providing welfare and conducting research. A typical overriding goal is one of self-maintenance—one of the ways it achieves this goal is by replicating its beliefs and practices. This may happen in a more or less predatory fashion, and in part takes place through communicative practices which have the intended effect of norming participating agents—of fostering adherence to beliefs and practices. In a general sense, having goals gives a culture a quality of intentionality—its practices are directed and goal-oriented, much as those of a biological agent might be. Retaining the organistic analogy, cultures reproduce, evolve, inhabit spaces, communicate with other cultures, and ultimately expire. While analysable and modelable, this cultural activity is partly stochastic, predictable only within broader, non-deterministic and probabilistic parameters.

A culture also operates within a general environment—what it sees as its ‘world’; or, in Habermasian terms, its ‘lifeworld’ (Lebenswelt). This environment supports other cultures; cultures can stand in structural relations to one another. Cultures can even be nested; for instance, when a greater culture harbours an embedded revolutionary cell. The relationships within and between cultures constitute semi-porous, permeable networks—sub-cultures, cross-cultures or ‘hybrid cultures’ are all examples. These structural delineations and permeations can be traced through the practices enacted within those cultures—canonically, within discursive practices. Discursive practices produce epistemic artefacts—representations of knowledge—which reflect the perspectival orientation of the culture towards the objects it encounters—or engenders—in its world. However, a perspective is not fixed—it reflects a point-in-time reification of a floating, variable conceptual scheme, which coordinates the production of beliefs within a culture. Other forms of practice, discursive or otherwise, are always ‘kicking back’ against a given perspectival view, which survives just so long as it can withstand or absorb these challenges. This is particularly the case in ‘experimental’ cultures such as scientific and, in the narrow sense exploited here, ‘ontological’ ones—cultures whose a priori rather than by-product function is the very production of knowledge. A characteristic of such cultures is their own self-explication of the beliefs and practices they engage in—formalised in rule-governing theory and rule-governed methodology respectively. This characteristic ensures repeatable observations—a kind of perspectival continuity across time, space and other cultural boundaries. Perspectives also, critically, remain one-sided; from any point of view there is always another, perhaps infinitely many other points of view available, through other accultured lens. Aspects of objects are both seen and occluded under a given perspectival lens; belonging to a culture, no matter how highly self-reflexively critical, means sharing both its insights and its blindnesses (de Man 1983).

Cultures, then, have conceptual schemes or, in Quine’s other metaphor, a ‘web of beliefs’ (Quine 1964). These beliefs can be described as structured like a network, spanning from the concrete, synthetic and empirical through to the abstract, analytic and conceptual—some beliefs are closer to the world than others. Again following Quine’s breakdown of the synthetic-analytic divide, a belief can be plotted along a scale of ontological-epistemological: ontological—here in the philosophical sense—to the extent that it refers to objects in the world; or epistemological—to the extent that it refers to other beliefs (or their expression in language). A belief is canonically expressed in a proposition, an assertion of a relationship between concepts, objects and properties. Together beliefs are mutually supporting, forming in the ideal system a coherent, consistent and non-contradictory whole. Within the semi-bounded environment of a culture, contradictions may nevertheless emerge in discursive practice between agents. One of the roles of discourse is to establish the grounds on which such intersubjective inconsistencies arise, to make assonance out of shared cognitive dissonance. Collectively, a network or web of beliefs constitutes a perspective—or, to use other common optical metaphors, an outlook, a point of view, a vantage point or an orientation. A perspective, however, is not here a passive lens through which the world is viewed; rather it actively constitutes, constructs and intends—in the active, phenomenological sense of ‘intending’—how things are viewed and arranged. Actors partake or subscribe to belief networks to the extent they are imbricated in a culture, acquire its ‘habitus’, although this is never (quite) a total imbrication. Through the roles they play and practices they enact, actors rather develop more or less intensive, comprehensive and enduring commitments to a set of beliefs. Understanding epistemic extent—the degree to which a belief is taken to be knowledge—is an important part of developing a profile of a culture.

Beliefs are transmitted in language, via what Habermas terms communicative practices. Communicative practice generally serves to break down what are otherwise incommensurable divides between cultures, and permit actors to participate in the ‘game of giving and asking for reasons’, in Brandom’s phrase. Hence assertoric utterances are paradigmatic instances of communicative practices—used to proclaim, query, test, revise, transmit, reconcile, and, in part, maintain cultural boundaries and integrity. Conveying of beliefs in language, while bearing the risk of a dissenting response, is above all an economic decision—it results in less work for belief transmission than other kinds of practices, of a presumably more coercive kind.

Beliefs form ‘webs’ in a less benign Quinean and more insidiously Foucauldian sense—as socially norming practices, discursive and otherwise (Foucault 1970). It might be possible in some cases to identify beliefs which are intrinsic and ‘core’ to a culture—those which motivate practices and subsidise ancillary beliefs, and which constitute the non-negotiable intransigent elements of a cultural ‘perspective’. These are likely to be those which are practically intractable to empirical or communicative challenge, since their invalidation threatens maintenance of the identity and boundaries of the prevailing culture. Belief revision is consequently largely a piecemeal affair, at least within the confines of a given culture, as the ‘carrying-out’ of practices and even the revision of certain beliefs can only take place while the remainder of the belief system remains relatively stable. In this model, epistemic revolutions, as opposed to revisions, are rare.

Yet beliefs, as purely ideational and immaterial constructs, are essentially unknowable directly, and only can be inferred via the evidence of a culture’s practices. For ‘ontological’ cultures—those that produce explicit representations of some slice of the world they are concerned with—a conceptual scheme can be directly interpreted in highly regulated ways, via the semantic specifications embedded in those representations. Such explicit representations cannot, however, be interpreted, purely and unreservedly, as the accomplished perspective of the culture which produces them. Rather they are narrow, restricted and temporarily discrete frames on an ever-changing flux of objects—or, in another formulation, sense-data ‘perceivings’ which only a posteriori congeal into the sorts of things conceptualisable as objects—subject to a continually changing dynamic process of theorisation and practice. Moreover they are also products with intended communicative effects. What is made explicit, then—the arrangement of concepts—needs interpreting not only in terms of its mode of explication, but against what remains tacit—a broader background of cultural beliefs and practices. Unlike the first kind of interpretation, operating directly on the axioms of a system itself and proceeding along set or model-theoretic—in other words, strictly analytical—lines, the second kind is necessarily dependent on exploratory, heuristic interpretive devices, using suggestive rather than direct forms of evidence. It is possible, then, to present knowledge systems generated by a culture as, instead of a stark unmediated delineation, a sort of highly detailed foreground, cast against a vague, impressionistic yet significant cultural backdrop. The resulting ‘portrait’ of a culture is then comparable with other depictions—partly in the precise quantitative sense of two geometric conceptual graphs being compared isomorphically, but also in the deliberately imprecise sense of two holistic images being comparing impressionistically.

Picking up methodological cues from Gardenfors (2000), to generate a portrait involves analysing cultures across any number of possible dimensions, some of which might be especially salient within a particular translation situational context. Generically, cultures can be described in a number of commonly occurring dimensions: size, rate of growth, ‘core’ or foundational beliefs, practices, perspectives, material and environmental conditions, influence, aggressiveness, health, longevity, maturity, internal organisational structure, relation to other cultures, organisational type and purpose (economic, political, legal), and so on. Other, specific variables relating to the situation in which translation takes place can be used as well. One possible formalisation of generic variables is described in the set of dimensions outlined below. Regardless of exactly which set of variables are selected, and how they are respectively weighted for saliency, what matters at this stage is that it is possible to describe, qualitatively and quantitatively, the explicit conceptualisations and tacit structures, beliefs and practices which underpin them, as a kind of portrait or profile. Commensurability of cultures involves, then, a comparison of the quantitative and qualitative profiles developed in this way. The ‘tacit’ part of a profile is not, of course, truly silent—it represents aspects of a culture which need hermeneutic or heuristic interpretation, typically kinds of discursive, textual practices in which conceptual commitments need to be ‘drawn out’ and inferred. Hence the methodological strictures about what counts as evidence, and what limits apply to the inferences drawn from it—this form of interpretation is necessarily suggestive and exploratory, rather than definitive and explanatory.

Presenting a framework for commensurability

The above characterisation is sufficiently abstract to describe the vague kinds of entities which reside behind knowledge systems. The remainder of the chapter now makes a sharp transition from theory to practice, from a theoretical model to a framework that might help an analyst work through practical problems of system translation. As indicated earlier, treatment of differences in knowledge systems takes place at least implicitly in several common information technology tasks—system integration, database design, information retrieval, decision support, resource planning, project management, and so on. A starting point for the framework is to describe what might be an idealised translation scenario, to serve as an approximation of the various real-world situations in which translation takes place. This provides a way of orienting the question of commensurability from the point of view of an analyst engaged in a translation process. The formalisation of the model also provides a way of moving from a qualitative to a quantitative characterisation of commensurability. From here, several generic dimensions for describing knowledge systems and their underlying cultures are proposed. A schematic procedure for applying the dimensions is then discussed, followed by some concluding notes on the interpretation of the commensurability assessment.

It is worth briefly reviewing the motivations for the framework. Referring back to Chapter 3, ‘The meanng of meaning’, there are broadly two ways of handling differences in knowledge systems. Computer science approaches focus on how to achieve individual concept alignment. They typically employ algorithms and external data definitions to match concepts from different systems. Matches can then be used to develop transformation rules to convert data from one system to another. These approaches can be broadly described as forms of semantic atomism—concepts are primary to the schemes containing them. Here by contrast system commensurability is considered in terms of plausible schematic alignment. This approach is fundamentally reliant on an interrogation of the cultural character of these systems. Following one of Brandom’s key distinctions it can be considered a form of semantic holism, where the overall underlying cultural conceptual scheme is primary with respect to the individual concepts stipulated within it (Brandom 2000). Furthermore it can be described as predominantly interested in the pragmatic character of knowledge systems—what kinds of use they are put to. The difference in approach, then, is largely one of orientation and method; semantic holism as advocated here is consistent with algorithmic matching techniques described above, and it can be used as a supplementary heuristic to these techniques.

Applying the framework

Against the background of a broad charaterisation of the social web of knowledge earlier in this book, here we have proceeded to develop a general theoretical rubric and detailed framework for assessing the commensurability of both formal systems and the cultures responsible for them. The framework has four components: a model of an idealised commensurability scenario, a series of dimensions conforming to the demands of the procedure, a means for quantifying commensurability and a method for applying the framework and interpreting the results. The framework mobiles a series of analytic tools for understanding both the explicit and implicit commitments entailed by formal knowledge systems—those directly stipulated in the systems themselves, and those inferred through an examination of the background cultures in which they are produced and used. The series of dimensions uses the distinction between intrinsic and extrinsic dimensions to capture each of these types of commitments. The separation of the model and methodology permits the adoption of entirely different dimensions, allowing for considerable flexibility in how other criteria and even different ontological and epistemological assumptions come into play in the analysis.

Assessing the commensurability of systems is a necessary but ultimately insufficient step for building webs of meaning and knowledge around contemporary representations of knowledge. Assessment needs to be corroborated with specific pathways of concept transformation and translation; these, in turn, of course need to be guided by the kinds of methodological, heuristic and analytic rigour an assessment framework provides. However, inter-system translation threatens to become a task of exponential complexity as the number, size and underlying assumptions of these systems themselves proliferate. One way to corral such complexity—within arithmetically rather than geometrically growing bounds, at least—is to develop interlanguages—kinds of reference languages designed to guide translation between the micro-languages articulated (again, both explicitly and implicitly) within knowledge systems. The next chapter looks at the development of one such interlanguage, and how it helps with building the social web.

References