Fwd: Defeasible Logic

Opps - forgot to email to list...

Per below.

Timo.

---------- Forwarded message ---------
From: Timothy Holborn <timothy.holborn@gmail.com>
Date: Mon, 14 Mar 2022, 6:09 pm
Subject: Re: Defeasible Logic
To: Dave Raggett <dsr@w3.org>




On Thu, 10 Mar 2022, 10:02 pm Dave Raggett, <dsr@w3.org> wrote:

> Thanks for the pointer.
>

Np. Apologies for the lack of engagement so far...  I'm still working
through some puzzles, as may be inferred to some degree by my notes below.

I had a 'to do' item to post a note about 'logical english'[2], that i've
just done...

 It seems weaker than plausible reasoning, though, as the latter can cope
> with uncertain, incomplete and inconsistent knowledge using qualitative and
> quantitive metadata.
>

wiki entry =  https://en.wikipedia.org/wiki/Defeasible_reasoning although i
remember others..

Maybe part of the requirement is to look at how to support multiple
modalities & interop?

Do we have a list of different methods developed historically? I could
start a Google sheet...  but i'm not convinced, yet, that there's that many
of them...

The notion of “proof” as used in [1] is perhaps more mathematical than
> grounded in the uncertainty of real life where “truth” is a rather softer
> notion.
>

I think part of what they were trying to achieve was a way to describe the
logic of the conclusion, rather than suggesting the conclusion was
necessarily absolute. In some ways, I guess, perhaps there's ways to build
kindness into the ontological structures - which is shorthand, for a more
qualitative consideration...

Although your note stimulated a variety of considerations that I've made
some attempt to provide considerations about, below...  I guess, the
particular focus here - is about natural (biosphere) agents and the impact
of technologies - rather than use-cases relating to artificial entities /
agents (companies, software, etc.).


Re: Proof & truth, both have uncertainty, temporal & other considerations.

I've attached that image of the different species being judged by their
ability to climb a tree...   Point being, the qualities of the infosphere
[3] have various (impactful) distincts; therein also, considerations about
the implications for legal & socio-economic aspects.

I've got various diagrams here[4] but one of the diagrams that looks at
temporal considerations is this one:
https://miro.medium.com/max/1400/1*OD62QRiDXQd5mBdR_MiFpQ.png

The point of that diagram is about whether someone is judged based on a
claim or whether and/or how cyber-social systems may provide for
adjustments due to, say, temporal social faceting.

This might be a use case or user story exercise to more fully explore and
respond to with greater detail. Although, a question becomes about carriage
& intended beneficiaries? Perhaps that's part of the modality
considerations.

Therein also; a need for trusted (legal) agents.  Part of the historical
'human centric' stuff was about trying to define mechanisms where there
could always be a cost of someone going to prison as determined by a court
of law for wrong-doings - rather than other systems that seek to create
patterns where social acceptance of fault (causing injury to others) may
suggest its the fault of some software or a legal personality (company)
with no penalty (regardless of harmed put upon others as a consequence of
actions) upon natural persons...  Now, when I read WEF documents that
appear to be linked to my works on knowledge banking ecosystems, etc. (
https://www.weforum.org/reports/advancing-digital-agency-the-power-of-data-intermediaries
) these sorts of considerations appear to have been lost....  which gives
rise to the 'modality' consideration that should, at a minimum, be
labelled... IMO: there's differences between 'human identity', how others
perceive who you are as a being/entity and how systems to support that are
made to work; vs. a datacollar that issues an identifier used for others to
define you, regardless of coherence with reality - although, an agent of
manifest with a capacity to distort & define for purpose, which, if the
modal structures are undefined - becomes a form of definition in itself...

Some of the most impactful use cases (whether and/or regardless of how
they're understood at the time by the operator rather than the subjects)
there's alot of private & sensitive information that's likely involved and
has temporal aspects linked to it, so whilst agent-centric truths may be
temporally discernable, there's nuances around how one might want to define
concepts of "absolute truths"... and whilst that's not something that's
achievable in an array of use-cases, the manner through which comprehension
is labelled (ie category theory?) processed, decisions around how means of
appending records or associating records or how aspects of significance can
be referred to (even if private / sensitive) has a meaningful impact on
coherence vs. distortive functions, interoperably with a 'wave function'
that in-turn infers a problem about dual-use / modalities...  IMO.

The temporal aspects are of great significance, both forward & reverse to
present-time defined by the observer as an 'agent'?  (not sure of
'exactness' language, seems more important than i have time to clarify
right now).

NB: <http://goog_1806491889>https://www.sciencedirect.com/science/article/pii/S2772941922000011

(nb also: https://axon.cs.byu.edu/Dan/678/papers/Recurrent/Werbos.pdf )

broadly speaking - the modal structure i've been trying to address is what
i call #RealityCheckTech - which could be applied to help people ensure
they're aware of how much toilet paper they may need before engaging in
violent activities at a shopping mall, as occurred in 2020 in many places;
or indeed also, means to ensure there's an ability to annunciate
differences or nuances between facets of reality / fiction vs. non-fiction,
which in-turn influences behaviours & whether we build systems coherent
with a concept of cyber-peace-fair (fair dealings) infrastructure, or the
opposite. Where there's a void of infrastructure supportive of peaceful
behavioural engagements, one might imagine that void being filled with
alternatives - i can think of several examples over recent history - the
consequence thereafter becoming, whether it's possible to build tools for
peace ATM; and if there is a well-supported interest in doing so, then how
do we define structures or boundaries around how that work should be done,
what it should declare or include for purposes of safety - particularly
within the civilian domain; and as noted in an article i wrote back in 2008
about preserving 'freedom of thought'[5]; given,

“the distinction between reality and our knowledge of reality, between
reality and information, cannot be made” Anton Zeilinger
<https://www.nature.com/articles/438743a>

How do we make sure we're not producing tools that place biosphere entities
into various kinds of a 'databox' that may inhibit their ability to look at
'the adjacent possible' https://vimeo.com/130884499

Whilst the principles of 'the charter' (from a UN context) or democracy /
liberalised democracies - infers that human beings are in-control of
determination - the way software is increasingly made to work, appears to
have meaningful aspects on the realities of these philosophical ideals.
Some might consider that we have a fabric of pervasive surveillance where
an enormous amount of information is stored about our lived experiences,
somewhere - yet, there's a multitude of examples where people have needed
to make use of that information stored somewhere to protect life, dignity,
but the permissions frameworks applied upon our infosphere have 'directed'
that the information is not available, which links to the concept of
'causality' and indeed also, 'logic' or 'cognitive bias', etc.

Perhaps, part of the function could be to consider how to document the
ideological hierarchy / modal logic & then apply COG-AI to provide an
inferencing graph of the consequences applied upon agents, in association
to agent-centric ontological predicates?

Thereafter; when considering the concept of 'defensive logic', perhaps part
of the consideration was about circumstantial deliberations rather than
processes for absolute convictions...  idk.

Timo.



[2] https://lists.w3.org/Archives/Public/public-cogai/2022Mar/0003.html

[3] https://en.m.wikipedia.org/wiki/Infosphere
[4]
https://medium.com/webcivics/inforgs-the-collective-info-sphere-67a660516cfd
[5] https://www.webizen.net.au/about/executive-summary/preserving-the-freedom-to-think/




> On 10 Mar 2022, at 07:47, Timothy Holborn <timothy.holborn@gmail.com>
> wrote:
>
> Hi All,
>
> I was recently reminded of this work on Defeasible Logic[1][2]
>
> Thought i'd note it...
>
> [1] http://www.defeasible.org/
> [2] https://www.governatori.net/publications.html
>
>
>
> Dave Raggett <dsr@w3.org> http://www.w3.org/People/Raggett
> W3C Data Activity Lead & W3C champion for the Web of things
>
>
>
>
>