If we accept that the “substance” of a period is made of cultural phenomena, quite naturally these phenomena cover an irregular area in space-time. A period may slowly spread out from a kernel region, and still flourish at distant areas when it has vanished already at its origins. In other cases it may again “retreat” to its origins, or even separate spatially. The precise boundaries are naturally fuzzy, except may be for cases of modern political systems with precise geopolitical boundaries at any time and islands. Even then, the precise boundaries may be too complex for the utility of a general thesaurus. Therefore we foresee two kinds of approximation (fig.3) of absolute bounds. [...] The first (“max space-time”) serves rough orientation and narrowing down database searches for finds: The total of all places where the period flourished at any time is approximated by an outer (larger) spatial bound . This might be by the identifier of a geopolitical of geological unit, or by a geographical polygon, or both. The starting phase and the terminating phase when the overall period flourished is approximated by outer temporal bounds , expressed by two date ranges. Note, that last date for the begin might be after the first date for the end [...] The second (“Space time analysis”) serves a more analytical description, in particular for cases where the dates differ significantly from place to place. (fig.4) For this purpose, we foresee a break-down of the total area into individual subareas (fig 5 ), which can be associated with individual data ranges, in the same way as for the overall approximation. In general not all individual places, where a culture flourished, may be known. Our knowledge is incomplete or changes in time. Therefore the space- time analysis is normally an incomplete list of individual centers of a culture. Areas included in the “max space time” but not in the space-time analysis imply such areas for which a reliable assignment is unknown.

As its primary role is to enable exchange and integration between sources of cultural heritage, it does not go into the details necessary to create and manage a thesaurus of periods. It does not propose a standard format nor does it analyze how to define a period.

The CRM models periods as a generalization of events, which occur in space-time. It uses temporal and spatial relationships even for cases where absolute time and space are unknown. The notion of history is based on timelines of objects becoming manifest in events.

Archaeologists create terminology about periods in order to classify finds by a schema of relative chronology (periods) and to relate sequences of settlements and cultures. The use of a certain terminology implies a number of ideas that may differ from researcher to researcher, so that there is a tendency to define cultural periods of even the same culture differently and to use different terminology for them (e.g questions about the definition or even existence of the Early Minoan III period caused disagreements between archaeologists in the past –see Zons 1967, Warren 1984). Period names may represent the same or different concepts depending on the view or social group. Traditional names may be preserved, altered or redefined (Barber 1987) on the base of earlier literature. Even if an established terminology about periods exists, it sometimes operates against explanation and understanding. (see comments on tripartite system of Evans, made by Dickinson in Dickinson 1994) The nomenclature principles lead to ambiguity since a name may designate several concepts. The problem is similarly known from biological species definitions, and seems to be common to taxonomies that follow the evolution of science.

Identity criteria are types of reported phenomena or the kinds of their observable traces that determine the unity and identify the cultural continuity of a period. We distinguish an instance of a period from another by a defining, unique “property”. By this property, any event, part or phase of the period is recognized to belong to the same (existence) period.

The schema describes a single period concept as a single document (in the sequence called “period definition document”) that may link to other period definition documents and auxiliary concepts, such as bi bliography or gazetteer entries. A period thesaurus is built from a set of consistently linked period definitions.

The schema comprises multilingual terminology for the period, the period definition in the narrower sense, its analytical spatiotemporal extent and spatiotemporal relationships, characteristic attributes and cultural influences, metadata about the initial conception of the period and the document that first defines it (the definition document)

Style can be used to sort and distinguish spatiotemporal cultural groups – culture includes style (Hodder 1987) - by specific characteristics; through a specific shape or decoration, style represents the specific society, which produced it (Renfrew and Bahn 2001). In such cases, we regard the period to be defined by style. Hence the appearance of the respective style is the ultimate criterion for the associated spatiotemporal bounds, and not the assumed political system or whatever other phenomena a researcher may asso iate with the observation of this style. [...] In the sense of (Guarino and Welty 2002), we regard a notion like “style-period” as a “rigid” property bound to the whole existence of any of its instances. As such, it “carries an identity condition”, such as “having the same style throughout its duration and spread”. This approach allows us to objectify the discussion, if different opinions about a period are due to different definition or due to different stages of knowledge or interpretation

Contrary to “defining criteria”, general characteristics of a period can be documented as part of their description. General characteristics are not part of period definition because they cannot change or determine the existence and the identity of a period. There are types of charact eristics, such as technological activities, social-political structures, economy and trading, history of war activities (and sequences/results), patterns of settlements and belief systems, generally different aspects of material culture. They do not define, but simply describe, and interpret a period. They are cultural activities that develop and vary during time- cultural phases, products of past human

The archeological notion of cultural periods is based on cultural semantics. Chronological systems based on such periods are notoriously controversial, due to the complexity of the relationships between contextual phenomena and spatiotemporal values. Nevertheless, they play a key-role in archeological discussion. Several projects have attempted to define standard systems of periods for indexing databases, interpreting the spatiotemporal or stylistic inclusion as a “broader/narrower term” relationship based on ISO2788 (e.g. The J.Paul Getty Trust College Art Association 2000, AAT 2000). This approach is not expressive enough to capture the actual complexity, resulting in differences between chronological systems being able to be justified by the differences in the initial definition and subsequent (re)interpretation of observations

Spatial reasoning (through region-based theories that mirror interval-based theories) has been done by (Cohn, Bennett, Gooday and Gotts 1997). In addition, there is considerable research and development into spatiotemporal databases (e.g. Koubarakis, Sellis. et al. 2003) to study the evolution of ecological, economical or other phenomena on earth. Most reasoning in computer science restricted to calculating operations with temporal information, purely spatial or spatiotemporal, rather than semantics of contextual properties

The presented methodology does not aim at the prescription of correct scientific opinion: rather, it aims at the overview and combination of established, alternative opinions using a methodical, comprehensive approach.

Systems for chronological reasoning in archaeology typically concentrate on statistical approaches to date archeological strata. These approaches normally regard the studied phenomena as simple, well- defined and associated with precise points in time. The Perseus Digital Library (Smith 2002) uses spatiotemporal maps and in dexes to search for documents about events.

On one side, computer science concentrates on temporal reasoning in controlled industrial environments, typically using discrete events and time intervals, neglecting spatiotemporal dependencies (e.g. Theodoulidis and Loucopoulos 1991, Ligozat and Vila 1998, Cowley and Plexousakis 2000). Allen’s logic, for instance, is only temporal - not spatiotemporal. (In fact it is included in the CRM). Allen (A llen 1983) searches/refers to techniques (by describing relationships between temporal intervals) available for representing time knowledge in general applications as needed in Artificial Intelligence

This paper presents the development of a methodology for creating multilingual thesauri of period names based on the specifications of CIDOC, archaeological theory and results from Computer Science and Know ledge Representation.

The design of such thesauri contributes to: the basic problems of how to define cultural periods and their interrelations (even when precise time, space and other properties are debatable or even unknown), and to the applied problem of communication, that arises because of disagreements between scientists about the definition of periods (and generally on relative chronology) and the respective terminology, and their use in the literature.

As culture evolves gradually, respective characteristics increase and decrease in frequency and strength. This makes the definition of a period necessar ily fuzzy with respect to space- time, but it makes the periods themselves not less real or objective. Rather, spatiotemporal bounds make no sense beyond a certain degree of precision. Further, culture evolves in many directions at the same time. The degree of synchronization between different kinds of phenomena, such as style and political system, may vary considerably. This gives raise to multiple points of view. Necessarily, such views differ objectively in their spatiotemporal bounds. Finally, archaeological observation is restricted to the products and traces of some of these phenomena. The degree to which these are correlated to the non-observable phenomena, gives raise to subjective interpretations. Jean-Claude Gardin (Gardin 1990) analyzes this process in general terms.

The complex and multidimensional notion of cultural (historical or archaeological) period makes quite difficult a definition and a formal analysis of period and generally a relative chronological documentation.

We therefore propose to distinguish between the phenomena, that have left distinct traces and are taken as objective indicators for the coherence of the respective period (the identity criterion), and other characteristic phenomena, distinct or not, that are either product of interpretation or that are not directly associated with the coherence of the period as a whole. For instance, “Ming dynasty” is defined by the political system. Any change of our knowledge about the dates of rule of the Ming emperors will change the asserted temporal bounds of “Ming dynasty”. In contrast, Ming porcelain is a good indicator for dating finds from the Ming dynasty. However, change in our knowledge of the dates of production of a certain Ming porcelain style will not affect the temporal bounds of Ming, but at most characterize this style as no more central to Ming etc. In this case, the political system provides the identity criterion, and the porcelain production just a distinct, characteristic phenomenon of this period.

[...] thesauri are collections of terms with generally agreed-on semantics and not individual definitions or representations of a concept (Doerr and Kalomoirakis 2000). Well-designed thesauri of cultural periods should help to: - define cultural periods and phases based on the distinct characteristics of the archaeol ogical contexts that are used by the respective scientific community to identify their unity, rather than by time and place, - organize the relevant archaeological information about periods in a consistent and comprehensible form in a computerized environment, - classify museums objects w ith respect to chronology consistently, in particular between different geographical areas, - identify finds and assess chro nology in archaeological research and excavations - facilitate better communication and knowledge sharing about periods between ar chaeologists, through a standard representational structure.

[...] we identify a period concept with the term plus a reference to the first known publication that actually defined it, just as it is good practice in biology (Berendsohn 1995). We propose to create unique preferred terms as in biology (e.g.: ‘Fringilla coelebs Linnaeus, 1758’). If details of the original definition have undergone some evolution, we would refer to the most recent interpretation of it. This case is distinct from an actual redefinition, which would give raise to a new concept. In the case of diverging interpretations, we also propose to create two distinct definiti on documents. Any number of synonyms in any natural language may be added. Synonyms need not be unique. They guid e the user to the concept and preferred term. A URI (“universal resource identifier”) may serve as language-neutral, machine-processable identifier

So far, we have identified the following classification of identity criteria: style, socio-political system, technological capabilities, strata, and cultural influence. This list will be refined and extended in the near future.

It must be clear, that a thesaurus can never reflect absolute truth but only a historical stage of knowledge.

Intuitively, many systems regard points in time as a means to recognize events and the temporal end-points of periods, confusing definition with conclusion, and identity with non-unique properties (actually, the absolutely precise point in time when an event happens is not observable).

The data that relate terminology and concept definition support the classical thesaurus functions (Soergel 1996, Doerr and Fundulaki 1998):

• guidance for the user to find a concept in the thesaurus,

• clarification of its meaning,

• unique identifiers to be used for effective identification, indexing and classification of the (archaeological) material

• preservation and update of information, through a process of proposals/submission of new terms or of modifications on introduced terms. This process has been structured in another schema (fig 1). It is regarded as a necessary process since it contributes to the completeness and final edition (through a version control) of the period thesaurus (Doerr and Fundulaki 1998).

• a frame of an ontology for standardization of elements. (Doerr and Fundulaki 1998). An ontological analysis, which aims at supporting reasoning about the relations between periods, and periods and finds.

Definition is the most important information category. It is about defining every notion of period in a specialized way. “Cultural period” is a multidimensional concept about complex interrelations between cultural phenomena. The CIDOC CRM defines the basic notion of a period as: “"This class comprises sets of coherent phenomena or cultural manifestations bounded in time and space. It is the social or physical coherence of these phenomena that identify an E4 Period and not the associated spatio-temporal bounds. These bounds are a mere approximation of the actual process of growth, spread and retreat. Consequently, different periods can overlap and coexist in time and space, such as when a nomadic culture exists in the same area as a sedentary culture…". (Crofts, Doerr, G ill, Stead and Stiff 2004). This definition is based on a notion of coherence, a “unity criterion” in the sense of (Guarino and Welty 2002), that relates phenomena such as people following a certain life-style, political decisions, an economy etc. Th e CRM definition goes actually beyond cultural periods in the narr ower sense, but we shall adopt this definition here, intuitively restricting it to “cultural” periods.

Some events are closely related to the definition of a period. We mark them as “starting event” or “terminating event”. A frequent intuition is that some kind of distinct event should be responsible for the upcoming of a new cultural phase (as required by Biers 1992). The question is: can a single historical, religious, military, political or physical event have a definitive affect on a period? Did, for example, an earthquake or a volcano eruption as the one that happened in Santorini determine the specific period? We regard that an event may be one of the catalytic factors to social change and thus be loosely synchronized with the end points of a period. Only the rare cases of complete extinction actually terminate a period in the absolute sense. Generally however, we do not regard those events as causal to the properties of a period, and the change of a period may quite well happen without such an event. Th erefore we use these events as chronological markers, “meetings” (Doerr, Plexousakis, Kopaka and Bekiari 2004) of items, objects and persons that were present in states/events, rather than as part of the definition. Such events are not regarded to be creating types of periods in the sense of cultural phenomena.