Howard Burrows and and I will ask if we should move beyond the semantic web to a web of knowledge that supports claims and evidence. This came up in 2008 when SKOS was being finalized: there seemed to be support for a new complementary standard, call it SCOS, a Simple Claims Organization System. To do
linked open data or any ontology, don't we need to justify that
we have "data" (i.e., that they could be used as evidence
somewhere) and that the domain terms selected aren't biasing
the discussion?
1. BEYOND SEMANTICS: SUPPORTING A WEB OF SENSE AND REASON Beyond Semantics Cambridge Semantic Web, MIT 4/22/11 1 Howard Burrows Valeska O’Leary-Whitney
2. First research 1965: NH oyster studyPhD 1978: Physiology and Biophysics15 yrs bench neuroscience: UVM, Dallas, NIH15 yrs digital libraries/epublishing: NASA15 yrs philosophy of science 4/22/11 Cambridge Semantic Web, MIT 2 HOWARD BURROWS
9. 4/22/11 Cambridge Semantic Web, MIT 9 It was a hamburger, not a cookie. It was Jim, not Tom. Tom took the cookie! Tom wasn’t here, he was in Kansas. Tom baked it, he didn’t take it.
10. 4/22/11 Cambridge Semantic Web, MIT 10 SKOS SCOS Access to documents and data sets Access to answers and explanations Vision Data Information Knowledge Wisdom Question Hypothesis Data Answer Philosophy
11. 4/22/11 Cambridge Semantic Web, MIT 11 CALL FORACTION Howard Burrows ghburrows@comcast.net @basicSci Valeska O’Leary-Whitney valeska@azigo.com @valeskaux
Editor's Notes
Semantics lets us “understand” word usage, but is not even in the realm of “knowledge”. As an example, I hope my concern with the use of the word “knowledge” in SKOS is *not* just semantics. We need to move to the next phase of web development that lets us catalog statements not by the concepts they relate to, but by whether the statements are true or not.How do we make sense of what we sense, and come up with reasonable statements and actions? What changes to web practices might facilitate knowledge development?
45 years in research! From subcellular/subsynaptic studies of release and reuptake of neurotransmitters, through tissue culture studies of NASA astronaut neuromuscular junction, to positron emission tomography of waking human brains, I did a bunch of stuff in neuroscience and medicine (killed a bunch of rats!). Then also did a bunch of field studies in Earth science before moving to science communications, economics, and process.
This talk began at this Semantic Web Meetup last month. Both the science guy and the designer need the web to support discussion about what evidence warrants claims.
So if our motivation is as honorable as encouraging smarter decisions…Is there an opportunity to recruit information technologies to this end?I don’t like limiting this to “science”; I think we want to support discussion and decisions. How do we come up with all the best ideas, and how do we rank them so as to select the very best one in each context of action?
What we want to do is talk about whether and how one should propose and build a new web standard
In Fall of 2008, as SKOS was in final stages of acceptance as W3C Recommendation, I participated in discussions across multiple continents about a complementary standard that might actually address “knowledge” itself, rather than concepts in documents that “contain” knowledge.It was a fairly highbrow discussion to the effect that “We should do that later.” SKOS is already a term of practice in the library community and we need their buy in.So, is now a good time to do it? What might be involved? Suppose we called it SCOS and designed it to support discussions about “claims” or “hypotheses”…
A simple claim.
How do we capture persuasive data? How can we elaborate a term or a statement into a huge family of statements worth considering?
Amplify and justify statements or actions, convictions.This isn’t Kansas any more, Toto. Silver Light is putting 10,000 images on a screen; NASA Earth science data sets are terabytes in size. Perhaps we should move beyond links and nodes?Perhaps to “spaces”?
Two visions. Two philosophies.Traditionally, there was thought to be a pyramid with data at the bottom analyzed to information, applied as knowledge, and improved to wisdom. It was a linear path beginning with data.That was in the 1890’s.These days we know that we don’t count many of our observations as “data”. The term “data” Latin for “given” isn’t even right, rather we excerpt (wrench) data out of context for rhetorical purpose in defending our Point of View (usually in science, to defend our pet hypothesis; to get a paper past our buddies on the peer review team)Most advances begin with an “ow” or an “oh! What’s that?” From there we synthesize wild conjectures about what’s up, and then excerpt away, or even contrive to set up just the outrageous conditions that would support our favorite idea.
Make RDF change from Resource Description to Research Discussion Framework!How do we begin a project like this? What would come next? Would it be a reasonable working group topic? For now, I will try to use the HCLS Science Discourse WG.