2008-06-11 email for archive - Bjoern re: Information artifact vs entity

[Alan Ruttenberg]

---------- Forwarded message ----------
From: Bjoern Peters <bpe...@liai.org>
Date: Wed, Jun 11, 2008 at 5:31 PM
Subject: Re: Reminder: First Information Entity Ontology Workshop, this Monday June 9, 2008, 9AM
To: Barry Smith <phis...@buffalo.edu>
Cc: Darren Natale <da...@georgetown.edu>, Alan Ruttenberg <alanrut...@gmail.com>, Rees Jonathan <j...@creativecommons.org>, Tom Knight <t...@csail.mit.edu>, g...@mit.edu, Fabian Neuhaus <fneu...@web.de>


I really like the limitation of scope to 'information artifact'. This is also the way I had started to remodel the sequence ontology.  'DNA sequence information' as stored in GenBank is an information artifact. This information 'is about'  the physical primary structure of a DNA molecule. The fact that DNA molecules can be transcribed (biologically) to produce RNA molecules by the molecular machinery of a cell does not require humans to be around. Doing this gets rid of the current confusion in SO between things that happen with physical DNA, and things I can do with DNA information in my computer (gapped sequences, BLAST alignments, phylogenetic trees, etc.).

- Bjoern

Barry Smith wrote:

At 09:45 AM 6/10/2008, Darren Natale wrote:

Some thoughts and questions on DNA as information entity, mostly with respect to a number of inconsistencies that I perceive about scope:

1) What is a machine?  It seems that in the scope statement, "machine" was restricted to the mechanical, ignoring the biological (see Websters definition, stating in part that a machine is "a living organism or one of its functional systems").  There is a common term in molecular biology: "molecular machine."

Alan and I ponder rebaptizing 'Information Entity Ontology' and calling it the 'Information Artifact Ontology' -- thus narrowing the focus, and ideally leaving open the issue of whether DNA molecules are carriers of information
This choice is motivated by the prime need at the moment, which is to support OBI, and to support the annotation of publications, results, databases, etc., all of which are information artifacts.

Some comments on Darren's comments:

2) Does a machine read?  Not if we insist on "anthrocentricity." However, here again, the common usage of that word does allow for machines (computers in this case) to read. It also allows for a molecular machine to read DNA. By the way, note that braille is read by a human in an analogous way as DNA is read by protein complexes.

I
 agree that there are all of these uses of 'read', 'information' etc. From the point of view defended by Alan and myself, they are, however, metaphorical. For you the type: [protein complex] reading DNA is a sibling of the type: [human] reading Russian prose; both are subtypes of the single type: reading. I think most people would find this odd, for instance in light of the fact that the domains of the respective reading relation would be so different.

3) language is used as a means to convey information. I would argue that DNA is a language just like any other--it is expressive (that is, the deoxyribonucleotide "letters" can be arranged to mean different things), and it is translatable into other languages (amino acid and ribonucleotide letters). If DNA is a language, then it must hold information.

I think this use of 'translatable' is metaphorical too (it is a bit like saying: tapas is sushi translated into Spanish).
I also think that we need a clearer formulation than ' DNA is a language'. DNA is a certain kind of chemical substance. A language, surely, cannot be made of molecules. A language is some sort of abstract framework within which (e.g.) messages are formulated.

4) Would you consider a blueprint to contain information?  A set of specifications?  How is DNA not like these?

One difference, important to us here: (wild-type) portions of DNA are not artifacts (i.e. they are not products of intentional action).

Here is a list of features of information as Alan and I see it:

1. amplification
2. communication, thus the potential for remote control, action at a distance
3. preservation over time
4. small occupation on space of possibilities
5. low noise ('perfect') copyability
6. formulated in a background language marked by
       6a, compositionality, which involves also a high degree of
       6b. free associability (providing I follow the rules of grammar (or analogous rules in music?) I have a gigantic degree of freedom to create new things which are still language)
       6c a distinction between the grammatical and the non-grammatical (between the well-formed and the ill-formed), resting on widespread consensus
       6d substitutability of synonyms (thus an idea of shared meaning -- whereby it is meaning which gets communicated)
7. aboutness (every information artifact is intended to be about something)
Optional criteria:
8. potential for logical complexity; narrative objects (e.g. journal articles) always have 8.
9. potential for normativity (some information artifacts convey oughtness, also governed by logic and aboutness)

Interestingly, most types of Shannon-Weaver information, and bird songs, do not count as information because they fall short already on 5. and 6.
When we say that primates do not have command of language we are addressing something like their inability to master anything like a language satisfying 6.
How is it with DNA? Is there anything like mastery, or freedom to create new strings by composition, in the world of DNA (I mean in the world of DNA as it is before human beings start to create DNA artifacts)?
Photographs lack 6. I think the same goes for pixellated images.
Absolute music and dance lack 7.

Proposal: Information artifacts have to be bearers of information in the sense which involves satisfaction of at least criteria 1-7. To be information artifacts in our sense, databases have to be ABOUT something (a string of random numbers is not an information artifact)

The following are information artifacts in this sense:

serial number
batch number
grant number
person number
name
address
email address
URI
protocol
lab note
ontology
gene list
publication
result
license
document granting permission
contract
novel
textbook
newspaper
timetable
recipe
map
objective specification