Integrative Levels Classification project scheme monograph references

« Ontological approach


Knowledge characteristically proceeds by analyzing reality into simpler units. Although our experience consists of a complex set of heterogeneous stimuli and feelings, we as humans do not have any simultaneous intuition of all being as a whole. Some say that such holistic knowledge is God's privilege, others [eg Ran51] that it can be attained by mysticism.

The whole from which every single phenomenon can be distinct, and from which it is someway derived, is described by various words in different cultures: the Absolute, the Apeiron, the Logos, the Tao, the Brahman, the Dharmakāya, etc. In ILC, this original, unanalyzed whole is represented by an asterisk: * .

Apart from vaguely acknowledging its existence, we cannot say much of this whole without distinguishing more specific attributes in it. These are the concepts, which thus are the units of our knowledge [*Dah78; Dah81;89]. Each concept is expressed in each particular language by some terms, consisting of one word (earthquakes) or several words (even numbers, knowledge organization systems).

How can concepts be identified? In other words, how can we identify boundaries within the mass of perceived data? It is a task of knowledge organization to find effective ways to identify these boundaries, and to reflect them carefully in our systems.

The pragmatist approach emphasizes the influence of culture on this (a classical example being how the spectrum of visible light is divided into colour terms by different peoples), while the epistemological approach emphasizes the role of perception, cognition and intellect [*Gno09]. However, in our ontological approach, we assume that all this in turn ultimately depends on actual boundaries existing within reality. Otherwise, it would be difficult to explain why different cultures have independently classified the variety of phenomena like animal and plant species in very similar ways, with differences often related more to practical interest in them (eg bird species all have detailed names, while ant species do not) than to substantial alternatives in their interpretation [Dia66; Wil92].

Concepts can be defined, or at least described, by a set of characteristics [Dah74]: a girl is a young female human. Each of these in turn can be a concept (a female is an organism producing bigger gametes, able to give birth, etc.).

Demarcation between classes of similar phenomena is sometimes neat: chemical elements differ between each other for the number of protons in their nuclei, so that one more proton means a different element — hence the periodical system of elements is regarded as an excellent case of a natural classification [Hjø11]. In most cases, hovewer, things are more complex: classes of phenomena can only be identified by a set of characteristics, some even shared with other classes, but of which what is unique is their association. Exceptions are also possible: lizards differ from snakes for a set of characters usually including the presence of legs; still, reptiles exist such as the slow worm that have no legs but are classified as lizards anyway on the basis of all the other characters (legs having presumably been lost secondarily).

Another instructive case is the demarcation of dialects. It is well known how dialect changes continuously from any village to the next one; still, there are places where changes are so sudden and important that a border between dialect groups is clearly there. I have directly experienced this in a neat way while accompanying dialectologist Daniele Vitali in his recordings at the western borders of the Emilian dialect group. Speakers in each village translated Vitali's word list each time in only slightly different ways, but when we reached the village of Ottone everything was suddenly different, telling us that we had entered the Ligurian dialect area. A similar experience was recorded by Vitali at the train station of Castel San Giovanni, where an Emilian dialect is spoken very similar to that of Piacenza, 20 Km eastwards along the Po river. There he was astonished to hear a group of card players showing much more Lombard characteristics (unlike Emilian dialects, /a/ in ancient Latin open syllables had not shifted to /ɛ/, and negation was not redoubled); so he asked them, and learned that actually they were from the Lombard town of Broni, 13 Km westwards along the same river, confirming that that short distance included a major linguistic border, unlike the greater distance between Castel San Giovanni and Piacenza.

These examples show that natural borders between phenomena do exist. However, it is difficult to define a fixed set of characteristics (traits) defining the concept "Emilian dialects", as one or the other character of the set often disappears in peripheric regions of the Emilian area [Vit11]: the borders (isoglosses) referring to different characters do not all overlap. The only useful approach is then a systemic one, which considers many characters and finds that most isoglosses are parallel and group into a bundle, which can be thus identified as the border; a similar bundle of major isoglosses has been identified running across France from Bordeaux to the Alps, separating northern from southern French dialects [Cha80. P 111]. Thus classes are defined not by a single character but by a "particular combination", in the words of linguist Graziadio Isaia Ascoli [Goe10].

Such decomposition of concepts is sometimes reflected in natural languages, especially those with a small basic vocabulary, like creoles: for example, pianos are expressed in a creole language as "crying-boxes" [*Fos61]. This has suggested that all concepts could be reduced to a small set of primitive concepts, by which all others can be constructed. This idea is developed in Llull's and Leibniz's "ars combinatoria". However, it seems to be questionable whether we will ever be able to come to really primitive concepts, and how can we identify them. Concept combinations like that of crying boxes are metaphorical rather than definitional, and do not provide a satisfying complete account of what a piano is. While a KOS could express the concept of acorns as the combination "oak fruits" (ie as a faceted class), it should not express the concept of pianos as the combination "crying boxes": rather it should adopt a new class for pianos.

Also, primitivity should not be confused with parthood. Although it is true that a cell is made of molecules, or that a football team is made of individual players, defining them as bunches of molecules or players is not enough, as they also have their own emergent properties, which cannot be explained only as the sum of the properties of molecules or players.

The last examples suggest that a phenomenon (identified in knowledge by a concept) appears where the combination of simpler phenomena produces some new properties, like self-regulation, metabolism and reproduction, or harmony, line-up and strategy. These properties then are the characteristics of the newly identified phenomenon. In an evolutionary perspective, such interaction of pre-existing phenomena to originate a higher-level phenomenon has been called synergy [*Cor05-06]: eg eukaryotic cells are believed to have evolved from a symbiosis of simpler, prokaryotic cells of different kinds, now acting as functional parts (chloroplasts, mitochondria) of a new structure.

Phenomena can then be identified (and referred to by concepts) by observing the properties which make them different from other phenomena. Cells are a class of phenomena showing self-regulation, metabolism, reproduction etc., as opposed to molecules or castles which do not.

Properties typical of a large class of phenomena are described in knowledge organization as their facets. In turn, very general facets common to all phenomena are called categories. Thus characteristics, properties, facets, and categories all are features useful to identify phenomena at various degrees of generality.

In this sense, Poli observes that a domain of knowledge can be identified and delimited as a partition of reality where some categories can be applied [Pol10]: physics is the domain of entities for which it makes sense to speak about space and efficient cause; technology is the domain of entities for which it makes sense to speak about design and purpose; etc. This looks as a powerful approach to divide reality into non-arbitrary classes of phenomena (domains), which can form the main classes of our KOS. In turn, each main class can be further divided into more specific classes, again differing between each other by some more specific property (facet); and so on.

Division of concepts »


Thomas M. Dousa and Daniele Vitali contributed useful suggestions for improving this section.

References cited in this section

Cha80 = Dialectology / JK Chambers, P Trudgill – Cambridge university press : Cambridge : 1980

Dah74: Zur Theorie des Begriffs. P 14 / Ingetraut Dahlberg = International classification. 1 : 1974. 1. P 12-19

Dah81: A referent-oriented concept theory for Interconcept / Ingetraut Dahlberg = International classification. 8 : 1981. 1. P 142-151

Dah89: Concept and definition theory / Ingetraut Dahlberg = Classification theory in the computer age : conversations across the disciplines : proceedings from the conference: Albany (NY): November 18-19, 1988. P 12-24 – State university of New York. School of Information science and policy : Albany : 1989

Dia66: Zoological classification system of a primitive people / Jared M Diamond = Science. 151 : 1966. P 1102-1104

Goe10 = La concezione ascoliana del ladino e del franco-provenzale / Hans Goebl = Il pensiero di Graziadio Isaia Ascoli a cent'anni dalla scomparsa : convegno internazionale, Gorizia-Udine, 3-5 maggio 2007. P 147-176 / Carla Marcato, Federico Vicario : ed – Società filologica friulana : Udine : 2010 | <>

Hjø11 = The philosophy of classification / Birger Hjorland, Eric Scerri, John Dupré = Knowledge organization. 38: 2011. 1. P 9-24

Pol10: Domain theory: a preliminary proposal / Roberto Poli = Paradigms and conceptual systems in knowledge organization : proceedings of the Eleventh international ISKO conference: 23-26 February 2010: Rome (Italy). P 145-151 / Claudio Gnoli, Fulvio Mazzocchi : ed – Ergon : Würzburg : 2010

Ran51: Classification and communication. P 179-181 / SR Ranganathan – University of Delhi : Delhi : 1951

Vit11 = personal email to the author / Daniele Vitali: 2011.02.17

Wil92: The diversity of life / Edward O Wilson – Penguin : London ; Belknap press : Cambridge (Mass') : 1992


Integrative Levels Classification. Philosophy. Division of concepts / Claudio Gnoli – ISKO Italy : <> : 2010.03.11 - 2011.07.29 -

  Integrative Levels Classification project scheme monograph references