10-08-2007, 04:33 PM
Decebalus/Andreas Gagelmann[/quote]
Decebalus states the problem with classification systems quite well. Where does one type end and one begin? Also Jasper notes that an image base can be first established, and the other ancillary data added later. Jasper is also right when he states there should be nothing that is overly complicated.
Mostly I used analytic methodologies in prehistoric projects. We lack anything like Greece or Rome here... so my work was done with lithics and pottery, which can get on your nerves quickly. I am not an expert in the Med by any means.
Dan, I believe, is more current than I am on what is most current in Archaeological thinking, so his comments would be most welcome.
The advantage to an analytic approach is avoid the problems quoted above. You get objectively defined clusters and more important sub-clusters of typologies without the subjective process of naming them first and wondering where one type begins and one ends. What stands out are regional and temporal anomalies and trends. These give you the best potential of research, if that purpose is the project intention. Once established the database can be accessed by research projects and graduate students for more advanced study, which is what they are for.
Once the classification schema is established, anyone who has basic familiarity can add to the database. The data is checked before final inclusion. Establishing the schema is what is difficult. The components have to be well thought out and stable. Doing this is requires submission of attribute elements, clear definition and justification of inclusion. This is established in an Attribute Dictionary, which gives typical examples. There is a tendency for these Dictionaries to get too large and complicated. The most often quantitative values that are successful seem to be this-to-that ratios.
An example, off the top of my head, in concept would be something like:
Material
1 = bronze
2 = iron
3 = other
Helmet skull shape:
1 = round
2 = pointed (height less than width)
3 = pointed (height greater than width)
4 = Lateral ridge
5 = Domed -- as in the later Corinthian. (This is an example of needing a clear definition)
Construction:
1 = one piece
2 = halved (lateral join)
3 = halved (lateral join) with additional components
Nasal
0 = none
1 = vestigial
2 = straight
3 = flared near nasal terminal
Neck Guard
0 = none
1 = short
2 = extended, flared at end
3 = extends to cover cheek -- as in a Corinthian
Cheek piece
0 = none
1 = prototypical ( short point)
2 = extended, (attached)
3 = extended, (separate) -- hinged, for example
4 = integral with neck guard -- as in an Argos or a Corinthian. (Note the Argos may be of multiple pieces.)
And so on as necessary, without being to complicated. Most of these projects that fail are because they get to be too detailed for no real reason.
A simple bronze cap would be: 111000. A late Assyrian would be: 231002.
An early Corinthian would be: 112234. Compare this to a later Corinthian which would be: 141334. The numerous variations on the Corinthian can be objectivily identified and classified along with frequency. Frequency is often an overlooked factor, as often, more attention is paid to variation than actual numbers of examples.
The attribute list would be more extensive than I give above, of course. A computer search would link close matches into clusters, i.e. objective affinities. Consider what can be concluded when you (the computer) get a cluster of six 113452244s (made up example) at three locations, but two of the locations have earlier dates than the third? What about a frequency cluster of near matches, say a group one or two attribute values off? Or something like that. This defines a Class or Type.
Locational information may be given by region. Exact locational information can be rather vague. For example, a grid on an X/Y axis, Western Crete would be something like K/D Eastern Crete, M/D. So three 113134 were located at L/D (Central Crete) and one, unpublished, at H/F (Southern Peloponnese). Most of my recent work is in GIS so I keep defaulting back to spatial referencing. Proximity indicating regional ineraction, in this case.
The descriptive information can be as Decebalus has given us in his "fact file" in his previous post. This gives previous catalog and textual reference.
Any member can add information that has access to the database, as you don't need any extensive professional training to do so, just familiarity with the process and the subject matter. Members who find examples can classify according to the Attribute Dictionary, and then submit them for checking and inclusion.
This may be more complicated than anyone wants to do. However, if established, it may open doors previously closed in terms of access, and give RAT a stronger academic standing. Also, the attributes can be given where a photo cannot because of copyrights.
My experience with setting these things up, is that in real life, there is often a lot of non-objective influences in the objectivity of the system. Attributes have been decided on by behind-the-scene personal popularity contests, the social movement of the month, love triangles, sobriety, political affiliation, and in one case I know of, a fist fight in the parking lot. None of the ladies involved were seriously hurt, BTW. Anyone who works in Archaeology for any length of time should be qualified to be a Soap Opera script writer. It is important to have oversight and justification of inclusion. One pottery class was established by the inclusion of a piece of linoleum as a joke.
I am just throwing this out as concept for now. I would have to ransack the U for more up to date case studies and methodology. My older texts may be a bit dated.
Ralph Izard
Quote:For some years I'm working on a database regarding all known Ancient helmets - a huge project you can imagine,... The main problems have been:
- Different nomenclature of types by English and German speaking scholars (for example the Hellenistic Attic type is often called Thracian in English literature, sometimes also the Phrygian and what is called Attic in English is known to German scholars as Italic-Chalkidian type and so on).
- A lot of mix-types with features of different main types (mainly in the Hellenistic section. For example the Phrygian-Boiotan, Attic-Boiotian Pilos, Phrygian-Chalkidian etc.)
s
- What can be called a "Greek helmet"?
The Italic-Chalkidian ("Attic") helmet is only found in Italy (there is also a Spanish variant) but is a clear descendant of the Greek Chalkidian helmet. The Apulo-Corinthian and Corinthian of Picenian form are also closely related to the Corinthian.
The earliest greek helmets have their origin in the Urnfield-culture resp. the Near East.
- For the main types of the late geometrical to early classical period there are already useable typologies but there is still no overview. The existing typologies have to be modiefied in some cases.
Decebalus/Andreas Gagelmann
Decebalus states the problem with classification systems quite well. Where does one type end and one begin? Also Jasper notes that an image base can be first established, and the other ancillary data added later. Jasper is also right when he states there should be nothing that is overly complicated.
Mostly I used analytic methodologies in prehistoric projects. We lack anything like Greece or Rome here... so my work was done with lithics and pottery, which can get on your nerves quickly. I am not an expert in the Med by any means.
Dan, I believe, is more current than I am on what is most current in Archaeological thinking, so his comments would be most welcome.
The advantage to an analytic approach is avoid the problems quoted above. You get objectively defined clusters and more important sub-clusters of typologies without the subjective process of naming them first and wondering where one type begins and one ends. What stands out are regional and temporal anomalies and trends. These give you the best potential of research, if that purpose is the project intention. Once established the database can be accessed by research projects and graduate students for more advanced study, which is what they are for.
Once the classification schema is established, anyone who has basic familiarity can add to the database. The data is checked before final inclusion. Establishing the schema is what is difficult. The components have to be well thought out and stable. Doing this is requires submission of attribute elements, clear definition and justification of inclusion. This is established in an Attribute Dictionary, which gives typical examples. There is a tendency for these Dictionaries to get too large and complicated. The most often quantitative values that are successful seem to be this-to-that ratios.
An example, off the top of my head, in concept would be something like:
Material
1 = bronze
2 = iron
3 = other
Helmet skull shape:
1 = round
2 = pointed (height less than width)
3 = pointed (height greater than width)
4 = Lateral ridge
5 = Domed -- as in the later Corinthian. (This is an example of needing a clear definition)
Construction:
1 = one piece
2 = halved (lateral join)
3 = halved (lateral join) with additional components
Nasal
0 = none
1 = vestigial
2 = straight
3 = flared near nasal terminal
Neck Guard
0 = none
1 = short
2 = extended, flared at end
3 = extends to cover cheek -- as in a Corinthian
Cheek piece
0 = none
1 = prototypical ( short point)
2 = extended, (attached)
3 = extended, (separate) -- hinged, for example
4 = integral with neck guard -- as in an Argos or a Corinthian. (Note the Argos may be of multiple pieces.)
And so on as necessary, without being to complicated. Most of these projects that fail are because they get to be too detailed for no real reason.
A simple bronze cap would be: 111000. A late Assyrian would be: 231002.
An early Corinthian would be: 112234. Compare this to a later Corinthian which would be: 141334. The numerous variations on the Corinthian can be objectivily identified and classified along with frequency. Frequency is often an overlooked factor, as often, more attention is paid to variation than actual numbers of examples.
The attribute list would be more extensive than I give above, of course. A computer search would link close matches into clusters, i.e. objective affinities. Consider what can be concluded when you (the computer) get a cluster of six 113452244s (made up example) at three locations, but two of the locations have earlier dates than the third? What about a frequency cluster of near matches, say a group one or two attribute values off? Or something like that. This defines a Class or Type.
Locational information may be given by region. Exact locational information can be rather vague. For example, a grid on an X/Y axis, Western Crete would be something like K/D Eastern Crete, M/D. So three 113134 were located at L/D (Central Crete) and one, unpublished, at H/F (Southern Peloponnese). Most of my recent work is in GIS so I keep defaulting back to spatial referencing. Proximity indicating regional ineraction, in this case.
The descriptive information can be as Decebalus has given us in his "fact file" in his previous post. This gives previous catalog and textual reference.
Any member can add information that has access to the database, as you don't need any extensive professional training to do so, just familiarity with the process and the subject matter. Members who find examples can classify according to the Attribute Dictionary, and then submit them for checking and inclusion.
This may be more complicated than anyone wants to do. However, if established, it may open doors previously closed in terms of access, and give RAT a stronger academic standing. Also, the attributes can be given where a photo cannot because of copyrights.
My experience with setting these things up, is that in real life, there is often a lot of non-objective influences in the objectivity of the system. Attributes have been decided on by behind-the-scene personal popularity contests, the social movement of the month, love triangles, sobriety, political affiliation, and in one case I know of, a fist fight in the parking lot. None of the ladies involved were seriously hurt, BTW. Anyone who works in Archaeology for any length of time should be qualified to be a Soap Opera script writer. It is important to have oversight and justification of inclusion. One pottery class was established by the inclusion of a piece of linoleum as a joke.
I am just throwing this out as concept for now. I would have to ransack the U for more up to date case studies and methodology. My older texts may be a bit dated.
Ralph Izard
