Re: contradiction and controversy - was: Is it necessary for the semantic web to be self contradictory? -

> On 19 Jun 2015, at 18:53, Larry Masinter <masinter@adobe.com> wrote:
> 
> I find myself having to remind myself in these discussions that the task of discovering what is ‘true’ about the world is much more difficult, intractable, and philosophical than the question of designing a practical application of artificial intelligence called the “Semantic Web”. 
> 
> If you move from asking “is X true?” to “do I believe X?” where every agent maintains its own level of belief, based in part on the agent’s understanding of the meaning of terms and trust of the source.
> 
> Then disagreement is modeled by different levels of belief. The model has to account for “well-established” and “prestigious” as terms of art as well…. the social aspects of establishing reputation and authority.

Indeed modelling meaning is one of those philosophical projects that has a long history. I think that understanding that history can help in building the semantic web applications - a bit like category theory - or the new Homotopty Type Theory [1] - can help in functional programming or in developing physical models [2]. ( Though the semantic web is a lot easier to understand than HoTT ).

But you are correct:  semantic web applications are not here to a priori determine truth. The semantic web should rather be seen as a new form of writing where the new pens are Web Apps which via a friendly user interface can write these relations to the web. This writing to the web forms writing acts
- similar John Searle's to speech acts but where HTTP verbs form the key types.  The user of these web apps can then explore what the world would look like if what was written were true. Or the user could decide that what was written is false and explore those possibility - in each case by just merging different graphs of information. In metaphysical interpretations of modal logics such as David Lewis [3] this can be put in terms of a philosophy of meaning where depending on what one believes the actual world is, a sentence can be true or false. It is true in some worlds and false in others.

So yes the user can make decisions as to what he believes, he can change his mind, and explore other things.

But what grounds this?

Well I think this is grounded in publication acts.  Not all publication of facts is publication of scientifically neutral facts. Most of it is information on which one can act. Such as for example the proposal for Adam to meet Bryian at 14:00 at MIT's Distributed Information Group offices. The publication of such a set of relations by Adam and the acknowledgement by Adam by the publication of such a relation too ( the details need to be worked out), now means that these two actors have engaged in  a minimal agreement between themselves of what the actual world should contain: namely that meeting. The non respect of that by one of the actors would get us back to the problem of contradiction discussed earlier: namely it would require each one to re-evalutate a number of things of which perhaps whether they could work together ( something that could be easily dismissed if they have worked together for a long time ). 

By allowing coordination of actions as well as beliefs (which are a form of action) these tools can not just be useful but it is a pre-requisit for them to function at all. It is the pragmatic consequences, in part determined by inferencial powers, which allows us to communicate. This is argued in detail by Robert Brandom [4], an analytic pragmatist .  Of course inferential power need not be automatised by computers. Humans reading texts have been able to make inferences as to what can happen without automation - other than those provided by a reading brain. 

Of course once this is set up, we don't just question everyting we find or else we would never advance. We tend to questions things when they don't work - ie when we find a contradiction between our expectations and reality. Eg the door does not open when we pull on it.  We rely on others and their statements to help us move along: most of what we believe is taught to us by others. But with so many others, we find that we must rely on institutions to guide us as argued very convincingly by Bruno Latour in "An Inquiry into Modes of Existence" .  What institutions we trust is partly determined for us in advance of our participation. We need to participate in the pre-existing instutiton of language to be able to communicate at all. And in the world of the browser, we must trust that institution to for example put the correct certificate authorities into the browser, and not to provide backdoors in the code.

There are ways a user could then also publish his institutional preferences so that those could be used by the software to determine trust routes. For example the user could select his country's business instutiton profile, which could link to all the country's businesses (directly or indirectly via local authorities) and all the similar ones existing abroad [5]. The browser could follow such links in a web of trust ( where the national country is the center for that user) when specifying trust logos sites reached by that user.

> 
> I’ve been very influenced lately by Schneier’s “Liars and Outliers” as a perspective that models trust. The “semantic web” protocol stack puts trust and context at the top; making them foundational helps with modeling controversy, ambiguity, and security.

Thanks for reminding me to read that.

Henry


[1] http://homotopytypetheory.org/book/ <http://homotopytypetheory.org/book/>
[2] for example just page though the book "differential cohomology in a cohesive topos"
   http://ncatlab.org/schreiber/show/differential+cohomology+in+a+cohesive+topos <http://ncatlab.org/schreiber/show/differential+cohomology+in+a+cohesive+topos>
[3] Eg his paper "Language and Languages" http://www.andrewmbailey.com/dkl/ <http://www.andrewmbailey.com/dkl/>
[4] for example in "Between Saying and Doing: Towards an Analytic Pragmatism" 
    Brandom was a student of Lewis but moved towards a pragmatic re-interpretation 
     http://www.pitt.edu/~rbrandom/ <http://www.pitt.edu/~rbrandom/> 
[5] I argued this in "WebID and eCommerce"
     http://bblfish.net/blog/2012/04/30/ <http://bblfish.net/blog/2012/04/30/>
[6] A lot of the thoughts above on writing are also inspired by Bernard Stiegler's online course on Plato.    


> 
> 
> On 5/26/15, 9:58 PM, "henry.story@bblfish.net <mailto:henry.story@bblfish.net>" <henry.story@bblfish.net <mailto:henry.story@bblfish.net>> wrote:
> 
> 
>> On 27 May 2015, at 03:14, Larry Masinter <masinter@adobe.com <mailto:masinter@adobe.com>> wrote:
>> 
>> Missing reference, what is the distinction before “Form” and “forms”?
>> But the Semantic Web is a web of utterances by distributed players which are all intrinsically ambiguous and highly likely to be full of contradictions and differences of opinion.
>> The main ‘invention’ of the web (and Gopher before it) was the distributed authority of having uncoordinated but linked data resources. A semantic WEB needs the same freedom for different serves to make contradictory assertions… is that what you mean?
> 
> I agree. 
> 
> What has not yet happened is specification of vocabulary for disagreement. Various philosophers in France, including Bernard Stiegler and Bruno Latour, have put a lot of emphasis lately on the importance of controversy.  
> 
> Even though a controversy could emerge from a pure logical contradiction such as the following example illustrated in the diagram below, where we imaging a well established ontology from a prestigious organisation that creates two disjoint categories of bosons and anti-bosons. Then we can imagine well respected scientists B that classifies an object as a Boson, and yet scientist H that then later classifies the
> same object as an Anti-Boson.
> 
> As he discovers this difference, H adds that the information on B's </thesis> resource is a falsehood.
> This is neither a Like nor a Dislike. It's both in some sense, as controversies can be very fruitful. At the minimum it is a pointer to reasoning engines that they should not merge both graphs.
> 
> 
> <Controverse-scientifique.jpeg>
> 
> 
> As is well know in logic everything follows from a contradiction: i.e. no distinctions can be made. And indeed once a contradiction is reached everything can be questions. In this case the controversy could question:
> 
>  * the ontology
>  * the statement by any of the scientists, or the procedures they used to reach those statements ( where they referring to the same entity? )
>  * the logical engines that showed these two as contradictory
>  * ...
> 
> A discussion on each side could bring to bear different points of views on the subject. At the end this could lead to a split of both communities, or a retraction of one of the three of some relation.
> 
> Not all controversies need be logical based. Someone could publish a relation such as 
> 
> <http://dbpedia.org/resource/Muha <http://dbpedia.org/resource/Muha>mmad>
>      <http://xmlns.com/foaf/0.1/depicti <http://xmlns.com/foaf/0.1/depicti>on> 
> <https://eagereyes.org/media/2010/empty-frame.jpg <https://eagereyes.org/media/2010/empty-frame.jpg>> .
> 
> And that could in certain circles be controversial. 
> 
> Henry
> 
>> 
>> Larry
>> —
>> http://larry.masinter.net <http://larry.masinter.net/>
>> 
>> 
>> 
>> http://masinter.blogspot.com/2014/11/ambiguity-semantic-web-speech-acts.html <http://masinter.blogspot.com/2014/11/ambiguity-semantic-web-speech-acts.html>
>> 
>> On 5/26/15, 1:59 PM, "Melvin Carvalho" <melvincarvalho@gmail.com <mailto:melvincarvalho@gmail.com>> wrote:
>> 
>> I was reading this quote lately:
>> 
>> In order for the Global Semantic System to be able to produce creative utterances, it is necessary that it be self-contradictory and that no Form of content exist, only forms of content
>> 
>> I was wondering if it applies also the semantic web and decentralization.  
> 
> Social Web Architect
> http://bblfish.net/ <http://bblfish.net/>
> <Controverse-scientifique.jpeg>

Social Web Architect
http://bblfish.net/

Received on Saturday, 20 June 2015 07:20:46 UTC