Today we are liveblogging from the Space and Time in the Digital Humanities Workshop, hosted by NeDiMAH and JISC which is taking place in London and follows on (in the same venue) from the JISC Geo meeting earlier in the week. As usual this is a liveblog so all of the same caveats as normal apply about errors and omissions. The hashtag for today is: #spacetimewg
Leif Isaksen (also of the PELAGIOS project) is introducing us to the day by saying that Greenwich is the place where space and particularly time is measured from. If you go out into Greenwich you will see a big laser in the sky and that’s the Greenwich Meridian. And if you look at Ptolomy’s Greek Parallels Intercept you will see that London is also marked there. Ptolomy’s regular grid was the first to start looking at time in
NeDiMAH is a new funded network for Digital Methods in Humanities and Arts from the European Science Foundation with objectives to create a map visualizing the use of digital research across Europe; an ontology of digital research methods; a collaborative interactive online forum for the European community of practioners active in the area. There are also a number of working groups.
Today’s event is arranged by Working Group 1: Space & Time coordinated by Jens Andresen, Shawn Day, Leif Isaksen, Eero Hyvonen, Eetu Makela, There will be four workshops over four years and you can find out more about these on our new website: http://spacetimewg.pbworks.com/.
The format for today will be four sessions on place, period, event, summary. We will have 30 minutes of position papers, then 30 minutes group discussions then 30 minutes of general discussion for each topic.
Panel Session on ‘Places’
What are Places? – Humphrey Southall, University of Portsmouth/Great British Historical GIS Project
My basic position paper is this slide: a table of different kinds of geographical entity and the role of gazeteers.
There’s a certain way in which places coincide with geographic features. In London a lot of our places have names like Royal Standard, Sun in the Sands, etc. So where does that come from? Well initially it’s a pub. What else is it? It’s in some sense a roundabout, a rotary. And again it’s a place, it’s a place on bus timetables. And it’s a conservation area – a names polygon with clearly defined boundary.
My second example is the Nag’s Head in Islington. A pub initially. Now an amusement arcade. But in Wikipedia the place is still there even without the pub. It’s a bus timetable location again. It’s also a town centre area: it’s a bounded polygon.
Of all these types of places the Elephant and Castle is the best known example but there are very many.
So if we look at another example. So we look at a map from the Guardian earlier this year of England’s most deprived areas. These are output areas from research though they are not
Jaywick (in Essex) was found to be the most deprived area in Britain. A discussion broke out in the comments about the second most deprived place, Breckfield. A commenter says that these areas do not exist, he is rebuffed by another who gives evidance: it’s a pub, its a centre, etc. It’s about a whole set of features.
Linda Hill writes about place with the example of Gruinard as a place. I disagree. It’s a series of features but in this example it is not social but geographical features. Historically this is how place is defined. Groome, for instance, describes Gruinard as “a bay, an island, a stream” etc. So we need to differentiate between geographical features and administrative areas or places. If we are interested in history and cultural research it is about administrative units not geographic features.
From Place to Facts – Franco Niccoletti
Lets go back to place. This is a very ambiguous term. It is an “extent in space (in the pure sense of physics)” and/or an abstract concept. We are not so interested in places but we are interested in facts and stories. We are interested in seeing which facts, objects and stories happen in a particular place. Relationships between places, between events, between objects are important to us.
There are some features and challenges of “place”. There is some fuzziness. Sometimes it is difficult to draw the borders of some place. Spatial entities, objects occupying space and located in places are affected by spatial relations – mereology, topology, is the place where an object is located part of it?
The expansion of the concept of place and the concept of appellation to facts is important here. A place X is identified by an appellation a(X) – which gives you an absolute reference, a relative to an overarching system. Or a relative reference, a relatvie to some local system. And most commonly by Place-name. We reason on appellations. We need to relate a(X) with a(Y) etc. to relate facts – what happened in the same place. However appellations have their own issues. They are imprecise. They are time-dependent – the place my change or the name of place may change over time. And they are also space dependent as there may be different appellations for the same place or the same appellations for different places. They are language and culturally biased. So is the fuzziness in appellations or places?
And finally we have Gazeteers and Thesauri. We are all familiar with these. Gazeteers are a list of appellations trying to normalize them, i.e. referring appellations to one another or to some reference appellation system (co-ordinates). But gazeteers do not take into account space, time or cultural variability and does normalising appellations influence our concept of places? Is there any way of dealing with needed extensions?
Place Reference Systems – Simon Scheider
Simon just finished his PhD at the University of Munst where I am based in the Semantic Interoperabiity Lab. We want to make information work together in terms of syntax and semantics. I will talk about an idea I had with Krzysztof Janowicz about place reference systems. I will talk about what reference systems are to explain what a place reference system could be.
Many of you will be aware of semantic technology, ontologies etc. Our opinion is that ontologies are very useful to do this in a certain way as they constrain options. But they have one serious problem which is that they do not account for the problem of reference. In philosophy this is about how to explain what symbols stand for. Reference systems in contrast account for this problem of reference in a practical way. Spatial reference systems do this well. There are formal theories – cartesian coordinates is a formal theory in a certain way. The theory of the primatives in the theory, the coordinates, are fixed in convention. There are certain operations that allow you to find a location from a certain location. There are other reference systems – calendars are temporal reference systems. We can reidentify those points in time using these reference systems. So how do we in general invent these reference systems. And a place reference system is desperately needed.
Places are not the same thing as locations. Every place has a location but they are not identical concepts. This is very clear from the previous speakers…
Comment from Humphrey: But not all places have a location…
Well where is the place of medicine in the Battle of Trafalgar. Is it the location on the battlefield where treatment took place. Or is it, say, the HMS Victory – but that is, at this time, in Portsmouth. It is a tricky issue.
So, the question for us is how do we generate these place reference systems.
There is much more to say on this – I wrote my PhD on this – but there is also a paper I would direct you to here: Place as media of containment by Simon Scheider & Krzysztof Janowicz [PDF can be downloaded here]
We are now breaking into discussion groups on Place.
The groups are reporting back:
Group 1 discussed the ideas of concept based systems of place. And also discussed patterns of movement and how that relates to these concepts. Behind all of that was the issue of addressing the needs of the specific user. It is more about customising the user interface to the user’s specific needs. That was the overall subject we were circling around on.
Comments from speakers:
Simon: on the user centered view. If you look at the issue of ontologies and semantics the user is always important. You should never leave this out of ontology or use of reference systems. One may come up with different ontologies, different reference systems dependent on the use of these systems.
Humphrey: In some ways what was key to my presentation was the idea of places in consciousness and discourse. Consciousness is very individual and not so helpful. Discourse is about sharing with others. So we cannot focus too narrowly on users, you need to focus on communities of use perhaps.
Franco: I may either totally agree or totally disagree with the idea of users! Which users you mean? Users of today? Users of tomorrow? Goals perhaps a better idea: what is the purpose of using this place, what do we want to achieve. Users can prioritise current use unhelpfully. We want to think about intended use, community of use and we can use shorthand of users. If on the contrary we mean lets investigate place and time for archeologists say then I would totally disagree.
Shawn Day: These are some great issues to raise! It’s really important to think these things through.
Laura: Thinking too much about uses can be problematic. Thinking about travelling say. We think about user travelling… if I’m a younger person I want discos and pubs. I’m travelling with children, so maybe I need hotels… these variant needs are important
Comments from speakers:
Simon: To address the user issue further. There is a difference between user centeredness and goal centeredness. We all have goals and we can share them of course. We can create ontologies that are widely usable. Goals and objectives can be shared.
Humphrey: The comment of talking about problems and not solutions… I was at a three day workshop in Seattle not unlike today. The problem was that that meeting does not seem to be leading to anything (other than a book). Perhaps a third of the people there agreed that linked data gazeteers were the way forward, the others didn’t know what to do with it. The PELAGIOS workshop do show the way forward in this area. There is work going on that needs some expository stuff.
Eero Hyvonen: I’m a computer scientist and from that perspective I want to give you some use cases. We have ontologies available but what we are looking for from those doing cultural heritage side: what are the problems you want to solve? Like those use cases about travelling etc. When we have those goals, those use cases, we can find issues and use those to find appropriate methods.
Leif: That’s certainly something we can talk about later today.
Group 3 discussed whether or not we are looking for a global solution or some local solutions for a problem. For instance archeological data structures of local grids, local reference systems could be referenced to a universal reference system on a use case or type basis might be better than trying to create everything. So if you have a book perhaps it uses a book system that refers back to a global system – maybe a way to deal with groups of things rather than a universal system.
Simon: That’s a very concrete problem. This is a problem we have in archeology, also in history etc. We have local very hard to understand systems. We need to understand and translate them back to other systems to understand thing. We need to think about what would be in that general reference system to solve this. A solution should also be triggered by practical questions and this is a very good one.
Humphrey: Is this about spatial coordinate systems or something else? If you start from an archeological systems you would think of location as fundamental but names much more fuzzy. If you define things by a name but without a location that can still be quite a solid thing. For instance if you take the example of Camelot. No-one knows where it is but it is a very clear, very concrete entity. And it is a geographic entity. But we do not know the location and in history this type of entity is not unusual.
Franco: Well possibly Camelot is not a historical place but all the same my presentation was about place as where things happen rather than location and Camelot would fit into this system. Space and place are strongly linked ideas but on the other hand using the same framework for very different content leads to very poor information. So for example libraries – general libraries are valuable but specialist libraries are also essential.
Comment: What is most useful gets used the most. Cross referencing systems can be dangerous in some sense. There is a Darwinian element here – what happens with systems that are not as useful.
Simon: Use cases are helpful. You can start doing something, see how it is used and understand the use case that way.
Group 4: We ran a little out of time discussing the issues. One thing that is worthwhile to add: we talked a lot about pragmatic approaches. We discussed that place is much more a social construct than a geographical thing so how do you establish equivelance. And when does it not matter to have equivalence. We also thought about PELAGIOS’ approach – mapping systems against a baseline gazeteer. People can annotate their own data then find connections between data. Solve the common denominator issue that enables you to achieve interlinking.
Humphrey: Again I’m not sure what our collective baseline. In some ways my presentation was a grossly simplified critique of gazeteers. A lot of national digital gazeteer providers are very much thinking about features but time creates all sorts of muddles. Wikipedia and DBPedia have a lot of knowledge in them but they are scary in terms of features. I’m not sure GeoNames are much better.
Shawn: We have three real issues we really need to address and go to. The issue of a variety of case studies is a good idea but how do we deal with the unknown user and be as strong as we want it to be. And how do we deal with abstracting and not abtracting too far. There is a practical way of doing this, of thinking of which questions we want to answer. The Stanford folks for instance have been talking about how to deal with users that need simple tools to work with, who need to quickly understand the issues. People have been grappling with this for centuries so I don’t think we will solve this today but we will have some great discussion.
Panel Session on ‘Periods’
History in Context Ceri Binding, Hypermedia Research Unit, University of Glamorgan
Although we have three distinct sessions today it’s impossible to entirely separate these concepts. Objects connect to events, events connect to places and they connect to periods.
Simple attribute assignments contain a lot of complexity and implicit semantics and lack flexibility. We need to be able to document the statements we are making and the provenance of those statements.
We have been using the CIDOC Conceptual Reference Model to create an event based nmodel rather than just attaching a date to an object.
Periodization lets us subdivide time and cronology lets us order ad understand events. We also want to classify a period in some way – monarchies, style, etc.
In early periodization & chronology we have Erastothees was a 3rd century BC Greek Scohlar who established the first Chronographia of Greek history. Ussher’s Chronology looked at ordering events. His work was rather overshadowed by the fact that he included – as others at this time did – an exact date and time for the “creation” 23rd October 4004 BC, noon. That sort of discredited his work but it can still be useful. If you see a passage from the Common English Bible – Luke 3 (referring to John the Baptist) gives a reference to a date in the rulership of a Roman emperor that we can cross reference with Roman record keeping to anchor events in a particular time.
So when we model periods with CIDOC CRM we do not need to fix the exact time span but we can connect relative timespans. We can assign attributes to an event that helps us explain where this assertion is made – multiple people can make such assertions and give multiple and conflicting definitions for a particular period of time. It’s important to have that multiple vocality in time periods.
We also need to understand period relationships – A is before B and B is before C, say. Putting periods in relative order is more important than having exact dates attached. So for instance we took English Heritage’s SKOS concepts and connected them to CIDOC CRM entities to build up conceptual entities.
We have also created a simple tool for looking at dates and time periods available at http://hypermedia.research.glam.ac.uk/kos/star/time-periods/ which we showed at the first PELAGIOS meeting earlier this year.
[APOLOGIES. Our Wifi connection died here and notes from the second speaker were lost. We will try to reconstruct these later today.]
Glauco Mantegari, University of Milano-Bicocca, Italy
[Notes destroyed by wifi issue so will follow]
Use of Periods in British Museum Documentation – Jonathan Whitson-Cloud
Why would we have a thesaurus? Well the British Museum’s purpose is world peace – the concept was that better understanding leads to equivelance and peace.
We have a more pragmatic set of reasons for needing a thesaurus. We have 1449 terms but not everything fits together oerfectly. We have all the usual partso o thesaurus terms. We try ot to use rlated terms for periods. We only use period/culture field to indicate production period. All of our use is very object orientated. We do not ave associations to other references. We have lots of ways to record uncertainty and fuzzy periods.
We don’t call it period but “material culture”. We really think of it a a cultural label rather than a time or place – we record that elsewhere. This information always interacts with other things. We do not include date but most commonly we add context through material, authority/regnal dates, production/person(s), school, state, ethnic group, find-spot, associated name or place or event. And some departments refuse to use periods at all.
Period is always part of a wider set, it always interacts with other information on the page.
Inference can be an issue – it’s always appealing to fill in as much information as possible. If a definition of a period changes you have to update lots of record. So we allow conflict. We have periods and we have dates and if they clash we allow that, it’s the reality of what we currently know.
What we like about periods is that they are conceptually simple – very good for a lot of our audiences.
We want our thesaurus to speak to others thesauri. We have made our British Museum data available as a SPARQL endpoint (here: http://collection.britishmuseum.org/Sparql/). Thesaurus could be extracted or referred to from that. We inlcude Periods which are freely share upon request and you could embed references (URIs) to other thesurus in BM etc. And we are keen to engage with projects like PELAGIOS.
And we are now discussing those presentations in our groups (and having some lunch).
Group 1 had some questions: how can you tell that the same time period is alike? Use cultural artefacts and cultural opinion – combining three approaches. how about using temporal techniques to define place. we also talked about simplifying models, how you can keep rich complex information but document it pragmatically.
Group 2: it is clear that objects are important. And objects connect to taxonomies and their own systems of organising. Context is crucial, both the object context and the museum context. These are concepts about grouping and differentiating between objects and time concepts. Some data is fairly fixed – archeological layers etc. We also talked about the user dimension. Library catalogues can be useful for users but museum catalogues have not generally been designed for use by users in the same way.
Group 3: Guaco talked about the vagueness of different granularities. We talked about different scales of ontologies etc. this morning. Is there a system to define the granularity of your ontology or your reference system? So that we can make better sense of granularity. And touching on Jonathan’s talk terminology is crucial. Just identifying key terminologies, taxonomies, ontologies in a given domain is critical. You need to be able to plug in some sort of controlled vocabulary or ontology etc. For me a big win for developing thesauri in RDF would be to find ways to cross reference them. We know that there is ambiguity but that’s where the Semantic Web vision does start to help understand those relationships and articulate the various meanings and systems in use. But I’m still struggling with where we connect spatial and temporal together. We can talk about the Romans for instance. We have a period but that’s only relevant to a certain cultural history in England. It means different things to people from other parts of Europe say. We need better spatial boundaries on periods for this sort of reason.
??: We had a litte chat after the panel session and we raised the issue of communality, communal understanding of concepts. And also about facilitating different types of users to make use of it. And on the technological side the claim was that this is far more advanced on spatial issue. But the panel table felt like the spatial community had some way to go on the conceptual level. So, comments on communality…
Jonathan: Communality would be nice but I think we it is more likely that we will share some things, we will define some things differently. And that should be fine. We need to assert useful knowledge. It’s useful to think of ourselves as a community. But we need to be aware of where we specialise both as humans and in terms of data.
Ceri: I think a baseline about a period, about a place. I think that is important to understanding, trusting and using data. An esperanto for our data here. Without that there is less chance of using data beyond what it was originally intended for. Interoperability is the promise of the semantic web but to get out of our silos we need to find those ways to trust and share. In terms of looking at periods and whether two are the same we need assertions and meaning not just labels when we compare these things.
Guaco: I agree with you. Of course the definition of periods and even the more general concept of time is very difficult to agree on. On a practical level and thinking through the practices of practices in some specific domains we do have some basic common assumptions on what period is, on what an event is. Perhaps it’s not formalised as models. But archeologists for example have a long tradition of defining that concept. Of course this new technology makes it possible to take care and represent some possible differences in meanings and to eventually let machines automatically understand different approaches. But I think too that basic agreements on concepts is needed otherwise it becomes very difficult to do anything. Thinking to ontologies, triples, RDF, anything… you need to attribute so these new technologies can help us to show the provenance and attributions for what you are asserting. This is a core concept in semantic web use anyway. If we have completely integrated information we need to know how to connect information, how to understand the reliability of some data versus another.
??: So we have an issue here. We have authority as a way to get communal understanding. Is this an accepted road to communal understanding?
Comment: Wikipedia has given us a new idea of community and authority. Encyclopedia Brittanica was top down but Wikipedia is about the community expanding and improving knowledge from the bottom up.
Guaco: Well Wikipedia as a model is not in contradiction to approaches discussed. Perhaps it is useful to consider periodisation beyond the community of experts. Often they have very specialised view but general users are also interested in this information and provide different views on the same thing. Important to interfaces and how users perceive these and use these terms in actual systems.
Comment: For example on Rotton Tomatoes film review sites they have a top critics rating a general critics rating and a user critics rating. And those might conflict heavily. One might argue that perception of some cultural periods might also share that conflict of opinions
Ceri: Some of those social softwares are very interesting and offer good opportunities. What I want is more than one point of view. Wikipedia doesn’t really solve that – where there is controversy the page represents compromise rather than multiple vocalities but yes, it would be good to have new voices.
Jonathan: There is a kind of race between Google and Wikipedia to deliver this kind of thing. Google are trying to make semantic search engine. They have quality but not quantity but they are working on it. The tools will come. More than one view is great though. Linked data really allows
Comment: You talked about attribution and provenance earlier. The way people use that data requires that provenance especially if the broader community is contributing here. I am interested in how we keep attribution metadata even when mashed up into new services. How do we make sure that attribution is retained in those end points.
Leif: It’s a bit off-topic… I know that attribution is very desirable. But I wonder whether when you take information and create secondary products you need to always attribute – the difference between a coffee table book and scholarly materials is significant in terms of attribution and citations. But you must make it easy to attribute data.
Eero: Google brought, a year ago, freebase, and then they are implementing that in the heart of the search engine and making use of this. One main issue there versus DBPedia. Freebase is created also by the public and selected by editors although volunteers, the main issue is to make it correct, always editors to try and keep it correct. Interesting to note. And another point to note. They do not use triple store in Freebase but a quatro store and that is to account for attribution to help with quality.
Nicola: There is a cultural consideration around attribution. One has to be careful not to use attribution in such a way that it implies credibility to a secondary use of the data. That is an issue in some other scholarly data contexts. If the original source data author is visibly credited it could imply endorsement and that could be create issues around sharing data.
Simon: I’m not sure that authority is the solution to this. It may help keep work to a certain standard but it does not address the heterogeneity of the data.
Panel Session on ‘Events’
Eero is introducing this session by saying that events are even more complicated than places and periods in some ways. We have so many aspects that are interested in the human sense of events. And from an interoperability point of view they are very complex to understand. So we have three different perspectives here.
Once Upon a Time: Space, time and event in modern storytelling – Laura Spinsanti, Spatial Data Infrastructure Unit, Insitute for Environment and Susainability, Joint Research Centre, European Commission
I’m not from the historical domain so I apologise in advance if I say anything silly about the past. So if we take “Once upon a time in a little cottage in the forest there was a little girl names Andrea…” we have a time and a place there. We then read about a dragon the forest – something interesting is going on. This is an event.
We have all sorts of new ways to tell stories. We have microblogging – people write stories of their everyday lives on Twitter. So for instance we have a story, a time and some indication of place with this brief stories.
HC Vent – Here by Dragons. From space to place and back – ancient maps tell a story, about important places (church, castle, coaching inns) and dangerous places (mermaid). GIS describes static reality – now it is perhaps more dynamic but it is a reality far away from what people can use, perhaps apart from scientists. And then we have neo-geography – the usable geographic information to describe reality. In some ways this takes us back to place and the activities we are doing in a place. A map can tell me about hills and rivers on a Leanardo De Vinci map. If I look at a proper modern geographic scientific map I need lots of information and skills to read that. And meanwhile through Twitter, through other social mapping people are creating these neo geographies for themselves.
Looking to time. Time is not subjective – there are clocks everywhere – and yet time is no more beaten by events – we live globally and there are many events at the same time. Time is also our modern obsession. The promise of neo geography is the idea that we can update our sense of place over time.
In a dictionary events are defined as “something that occurs in a certain place during a particular interval of time” but we are talking about the observed world and we are therefore talking about something important when we see something as an event.
And we have various sensing tools – EO sensors, VGI sensing which is social and participatory (and problematic). These sensors create a huge amounts of data – in our VGI project we collect from 6000 to 30,000 tweets per hour, noSQl DB, the cloud, distributed computing. We want to mine that data, we want the context and semantics around this data. And we have to deal with concerns about the importance of the event for the community – when we use data from social media we have a partial snapshot of that community and therefore a partial view of the imporatance and activity around that event.
So, in conclusion…
Is history written by the vistors? Well it is now more participatory, social, and more gender balanced perhaps if only in terms of perspective, from the bottom up.
But there is lots of challenge here. Imprecise, vague and fuzzy methods to use these new data sets. Time varying information needs to become a Standard Time representation; big data/scalability – a new scientific challenge; credibility – authority versus community. People are talking about areas they understand very well so they have and bring lots of their own context
Comment: I have a question… I kept thinking that you would show us the character in your story. why not? In normal narrative terms you would focus on the character.
Laura: I focus on space, time, and event but I could have…
Deducing Event Chronology from Narrative – Oyvind Eide
Holmen/Ore calculations work. We looked at documentation dated 1660 about a church being built, and another dated 1690 about it being built, and one from 1711 an account of the construction taking at least 6 years. Then an account of 1984 saying that a coin from a particular rein was found in the foundations. The idea is to reduce uncertainty. That’s fine for time…
Can we make a similar tool for spatial analysis? It is more complicated to move this into the spatial dimension. If you know something takes place in a larger area and that there are broader bounds for related events there may also be ways to reduce uncertainty. Based on my PhD project where I am modelling verbal and map-based expressions of geographical information I am looking at what is there in a textual description. You have to have leeway on connecting points when you look at, say, point A being a mile south of point B etc. and your true possibility room gets worse and worse as you expand the description. Is it possible to make a geographic dimension like that and actually reduce it down to make sense of the places by seeing that certain possibilities are not possible.
Narratology and how events are described – e.g. Bakhtin and his notion of the Chronotope – can help us understand the temporal and spatial relationships that are artistically expressed in literature. Maybe here we can understand where space time and narrative meet.
Nicola: How do you deal with the fact that the narratives you are comparing may actually be based upon each other – successive accounts building on those before. If you use those to verify each other that will surely be an issue in using this approach?
Oyvind Eide – It is important in the use of the system but I’m not sure about in design of the system. When archeaologists look at sources to evaualate a system so understanding sources here is very important in doing that
Comment: Have you done anything with regional connection calculus.
Simon: There are people working on spatial information and spatial relationships and they are trying to come up with a theory on this.
Laura: You can try to use contact information. If there is a description of a building, perhaps a building cannot be constructed in any place. Perhaps there is a building on a river so that you can exclude some possible locations say.
Oyvind Eide : For various reasons I try not to use pre-existing maps as many problems would be solved to quickly as I want to understand the uncertainty here.
Eero: I think the reasoning stuff is very important here. When we know that counties split and merge but not the population or coverage it is possible to reason with these sorts of approaches.
What is an event? – Ryan Shaw
I am an information scientist working with lots of historians, but a different sort of historians than most of those here – those working on recent twentieth century historians looking at radicalism and civil rights. I guess I see a difference between scheduled events and historical events that are more retropectively defined.
So I think we’ve actually gotten pretty good at modelling space and time and then we abstract these. But when we talk about space and time we actually want to model events and their possible relations. So what is an event shaped like. So according to Wikipedia events are shaped like a box – this is the source information for DBPedia, Freebase etc. So we have this near box with labels, participants, location, dates, etc. This is better than nothing but I think we can do better than this. But events are not necessarily blocks… maybe the Tetris form of events, the slotting together of many events.
Events do not have a specific shape, they shift and they are this mix that Humphrey talks about of conciousness and discourse. Our conciousness divides experience and stories into events that is to some extent culturally independent. But that same psychology is broken up further by language. Watching a movie, playing a game, those also trigger that breaking up of events as if we took part in these experiences.
So how people formulate event models in their mind follow dimensions of time; space; protagonists; causality; intentionality (Zwaan, Langston…) and how you store those events in your mind shifts depending on how you read a narrative. Many of these experiements have been with simplified texts of fairy stories. But looking at more complex events at any point of history you can tell a story at different granular levels. A story in days may not fill in detail of an account of an event by year. But there is a relationship between levels – you can tell the story of the 17th century can be told in centuries or in decades say.
There is an interesting relationship with place here. If we were planning a trip to the west coast we might say that we should visit San Francisco. But your itinerary between two different trips to the west coast may not coincide at all. So a Flickr map of San Francisco shows that tourists take totally different images to residents – there are two San Franciscos. In fact there are thousands. In events the same is true – there are thousands of different Arab Springs. There are a number of ways that each story constructs the same events differently.
So another example here is the Neighborhood project – Matt Chisholm & Ross Cohen (http://hood.theory.org/). Some stories have common paths and key tracks or recommended routes around an event occurs. So we can see a clear block like identity but that is built up over time rather than being inherantly true. Historians often have clear information about when things happen but they are often interested in disrupting that clear path of blocks. For example if you see a review of Blood Lands by Timothy Snyder there is consideration of how one can see a more broad view of Europe between Hitler and Stalin. When we build those structured paths and chronologies in our infrastructure of teaching history then we abstract those events.
What we are striving for is models of events where we can abstract between different granularities of an event. Through a shared level. Then a more nuanced view of a concensus pattern (e.g. the British WWII, the Japanese WWII). Then at the next level there is the individual narrative to compare different events. And what is interesting is how we can extract shared labels from these individual narratives.
Question: To what extent are you asbstracting the spatial out of your definition of events?
Ryan: I think that you can think about those different levels. If at the shared label level of World War II it is near impossible to make that terribly spatial. But at the individual narrative level that is going to be much more specific in terms of place. So there is a trade off between richly modeling events by location and protagonists, and abstracting away to the level of labels etc.
Simon: I think the granularity is the issue in terms of space as well as times. So is your approach a practical solution to model events?
Ryan: I can make it a little more concrete for the use case I am interested in, the history of the civil rights movement. So I have accounts of the movement. You can see the evolution over time of the scholarly accoun tof the movement and you can see how different sort of individuals record the movement. I am keen to identify local events and then aggregate ways for sharing models, shared events, etc.
And we are now going on to discuss this panel over coffee…
OK we have returned refreshed… the final portion of the day is:
Open Forum Discussion on Space Time – Lief Isaksen is chairing this
The hope is to identify common themes. To ask about anything important we’ve missed. And to discuss how NeDiMAH takes forward ontology here.
Methods and Technologies and Infrastructure are what we want to think about first. Both current best practice and current tools. And also what are the things that we need or could improve?
Methods – Current
Georeference is a good way to look at place. The Gazeteer is a footprint and is based on spatial reference but there needs to be an independent place reference system. That’s a theoretical issue.
There are a lot of things going on in building digital gazeteer. But these tend to be topographic mapping or they are crowdsourced but for use in historical projects. But there is a large potential for retro conversions of scholarly work like the Survey of English Place names, spatial authority lists etc. But those are expensive to do. But a good gazeteer is a big gazeteer but getting up to big properly backgrounded content is expensive and difficult. We need to consider that size isn’t everything. And we need to retroconvert historical and scholarly materials.
There are issues around clarity and IPR issues.
Need a vocabulary to link places to other places. We need other techniques not just gazeteers here. Place ontology.
We also need an alignnment of KOS.
We have gazeteers but there is more to ontology than gazeteers. We need a better formal theory and ontology of place.
Un-GIS – we have these concepts of locations and places that we can work with away from GIS.
We need a temporal GIS – a system that allows me to work with temporal boundaries or events as they change a place over time.
Response: there are some systems for these but they are not commercial and they do not deal with fuzziness or conceptual aspects. Secondo.
Comment back: I’m talking about something that does let me handle those concepts, those granularities.
If we are talking about software we use PostGres and PostGIS and you can use that for all that sort of data. It’s not a GIS but a relational database. A GIS is not the way to represent this stuff.
Response: you need something visual to work with that.
I would recommend looking at existing gaming and modelling technologies. Gaming engines perhaps useful here.
High quality metadata from mapping agencies, and for those to exist across borders. We need vector maps and vector quality. Especially for Scandinavia. And crucial from moving from name to place on a map.
Finnish land survey publishing everything open source btw.
There was talk of PostGres and PostGIS and you can in fact build visual elements on that through interfaces so a clear set of demands or requirements are the key part of making those technologies work here.
Validation of existing crowd sourced materials could be used to move towards a place of having reliable data that builds on that existing material.
Geo-parsing needs are very specific for historical materials and needs improvement. There is such an important element of context in parsing historical materials since places change over time and there is not only the need to create specific parsers for specific materials but also to have a way to understand the context of how those geoparsers handle a given placename at the time of that material’s creation.
Is there a need for temporal parsers as well?
We have frame semantic parsers.
We also have Freebase, dbpedia, geonames, pleiades etc. available to make use of.
We need to develop or improve parsers.
We need event parsers.
We need improved event gazetters
We need good event ontology
We do have CIDOC CRM (-EH)
The Linked Library Data (W3C) resources have some real usefulness as well.
And… if anything else occurs to you do email or contact or comment in the direction of NiDeMAH.
NiDeMAH is trying to think about formulating an ontology of methods. They did some work on this a few years ago on Digital Humanities but DH has expanded significantly since then.
I’ve been finding it useful to think about a cookbook approach – here is a method, here’s what it is intended to use, here are related methods, here’s an example, etc. And perhaps a way to see if that method is good or if you should look elsewhere. Not giving full information but full pointers for intelligent tourists can find out where to find out more.
Shawn: It’s a sensible process but it is a long term process. We’ve identified so many different approaches here. In the perfect world I’d love to send everyone home with homework. We are such a diverse group here and we may think that we are one group here but are we even speaking the same language? We have to see how people do use these things so we can find what that one big pain that needs resolving, what we actually need and what we actually mean by that term. We use ambiguous terms. Even if we collected one from each person here with narrative and process about why that is the big pain that would get us a good step down the line. That would be a great way to start to move the process forward.
Simon: So should everyone provide a use case?
Shawn: Yes, that would be fantastic. We will go through all the materials we are recording here today but if people can go into that bit more detail and elicit what those needs are that would be fantastic. If we can task that out to people that that would be fantastic. Even if it is just one use case each.
Comment: I think that it would be important to provide some form to fill in to help us to provide you with those use cases in a consistent way.
Comment: The cookbook approach sounds good. Each year the barriers to entry for this stuff get higher so I’m hoping for something more – step by step guides.
Leif: I am concerned about mission creep. About making things to vague to be useful. But we could look to provide a “You Will Need…” list to help explain the sorts of resources one will need to have on hand.
So… what will be happening next in this project? Well we will be writing our report for the end of January 2012. We will create some sort of wiki or forum to encourage people to contribute and comment. But we may ask you to target your knowledge and we certainly encourage you to engage in that process as much as you want.
We will also be planning our next workshop and will be in touch about that.
Finally this is our very first workshop so do you have any feedback about this event. Good or bad.
Comment: I really enjoyed all of the talks but now I know what you want to achieve it might have been useful to step away from the theoretical and look more at the pragmatic issues, the way in which issues are currently being addressed etc.
Yes, I understand that. You don’t want to be too pragmatic but we definitely take that on board.
Comment: There is always the issue of 100 flowers type thing. Balance needed between structured and unstructured?
Comment: Today wasn’t very structured and for a first event that feels right, perhaps the next one might be more structured.
Comment: It seems that there is no agreement on the spatial issue, theoretically time and space is the same thing. Can we be clearer on what we are talking about? Is there a case for making event centic say? What’s the balance between finite and infinite here.
I think this is a contentious space to be honest but maybe… is there concensus here. Perhaps there were issues in how we described the event today.
Comment: It’s terribly difficult to come up with a conception of space and time at this sort of workshop.
Comment: I agree… but if we don’t… who will? So for example if we treat them as two entities we treat them differently from treating them as one thing.
Comment: I think the way to broach that is to air the problems and find areas of commonality and shared issues and working on concepts to solve those problems. I don’t think that we want to chisel up the concept.
Do people feel it’s helped them personally in thinking through these issues and awareness raising today?
In terms of our report we will share and communicate our report and workshops will be on other issues of space and time – GIS, web mapping etc.
Eero: the next workshop will probably be in Hamburg at the DH conference. We can put out the theme list in call for papers and if we have proceedings of that workshop then there are already useful resources likely to come out of that.
Would you be interested in timelines, chronologies etc…
Comment: Well this is old news for me…
Shawn: But actually this is a disciplinary issue. So many new digital humanities people are entering this space and are new to this and we need to be able to give them some different expressions of these sorts of issues and ways into these areas.
How about GIS/Webmapping?
Comment: I don’t know, it would not be of interest to myself.
Comment: Generally about visualisation I probably struggle with that. I’d like to see a wide range of approaches. Specifically as it applies to space and time.
We do have another working group in this area so we don’t want to tread on toes… we need to balance what we do with the work the other groups are doing.
We will be putting out calls for papers, and communities will be brought together and you’ll hear about that as it moves forward.
Tomorrow is the Pelagios2 Hackfest (we’ll be liveblogging this). The idea is to explore open resources that are available related to history, culture, heritage using geography as a point of reference. Pelagios2 is based around the ancient world but actually the day is broader than that. We’ll have tech specialist and domain specialist and we’ll be coming up with quick wins and pain points in interlinking open heritage resources with geospatial concepts. We hope to find out what is easy, what is valuable and what can’t we do and why.
And with that we are closing the day with a giant thank you to all of the speakers, organisers and those recording the day.