On 1/10/2013 3:38 PM, Gillman, Daniel - BLS wrote:
> Ed,
>
> I don't agree with parts of your last paragraph -
> ""I disagree completely. Every representation of data is some kind of sign,
>yes. But the datum itself is the meaning of the sign. (Thank you, Frege,
>even if Peirce did get there first.)""
>
> Frank Farance and I define a datum as the association between a sign and a
>concept (its meaning) with a notion of equality defined for that concept. (01)
Fine. We disagree on the definition of 'datum'. Your definition makes
it a synonym for the ISO 1087 term "designation" (at least as formally
recast in the OMG SBVR specification, with the assistance of ISO TC37
experts). (02)
> As you no doubt will recall, equality is the one 'property' true of every
>datatype as defined in ISO/IEC 11404. I put the word property in single quotes
>to signify its reserved use in ISO/IEC 11404. If you leave out the
>computational part, there is no distinction between a datum and a word in
>natural language. (03)
I note carefully that ISO 11404 defines "equality" on "datatypes". It
certainly does not define "equality" on "concepts", which is way beyond
its scope. Further ISO 11404 defines a "datatype" to have a "value
space", which, for a concept would be its extension, or for the sign,
its denotation. I don't think ISO 11404 defines "datum", nor does it
define the relationship between "datatype" and "datum". It refers to a
member of a value space as a "value". So "datatypes" have "values", in
ISO 11404. It is silent on the meaning of the term "datum". (04)
Now, ISO 11404 was intended to integrate datatype concepts across
programming languages and databases and exchange files, etc. So, "if
you leave out the computational part", ISO 11404 is not clearly related
to the conversation. (05)
As to the relationship between a datum and a word in natural language,
that is not the same as the ISO 1087 "designation" idea -- the binding
of a concept (connotation) to a sign. A 'word' can have more than one
meaning, i.e., be associated with more than one concept, as any
dictionary will demonstrate. So, the way you defined 'datum' above, it
is not the same idea as a word in natural language. (06)
> If a datum is only a meaning, then what distinguishes it from information?
>And why do we have a representation for it? (07)
I simply don't understand these questions. I doubt that we agree on the
definition of "information"; and it seems to me that we have
representations for meanings in order to convey our intent to others. (08)
> The definition Frank and I have makes a distinction with information
>possible, and the representation is an essential part of a datum, which is
>consistent with practice everywhere. (09)
I agree that the computational usage of the term "data" often means
"machine-readable representation of information units", or perhaps
"machine-readable representation of information" (which is a higher
level of concern). I only aver that the meaning of the term 'data' in
science and statistics refers to observations -- propositions/facts
about individuals -- and it is not concerned with representation of
those observations. Of course, representation/expression is an
essential part of communicating information, but it is not a 'delimiting
characteristic' (ISO 1087 again) of the concept "datum" in those circles. (010)
So, your "everywhere" and my "everywhere" seem to be different
universes. If you and Frank are defining the term for some
computational purpose other than knowledge engineering, our universes
are different, and disagreement is to be expected. (011)
That said, I don't think your definition of "datum" as a "the
association between a sign and a concept (its meaning) with a notion of
equality defined for that concept" is clear enough to be useful. How
does it differ from ISO 1087 "designation" = association of a sign with
a concept? What is the meaning of 'a notion of equality defined for the
concept' as a delimiting factor on concepts? For example, is "country",
as a reference to a national entity, a datum? Surely we have a notion
of equality for countries. Is "kill", as a reference to terminating
life, a datum? Whether we have a 'notion of equality' for any verb
concept I couldn't say. This is why extending the ISO 11404 idea from
'value space' to 'concept' is problematic. (012)
-Ed (013)
>
> Yours,
> Dan
>
>
> Dan Gillman
> Bureau of Labor Statistics
> Office of Survey Methods Research
> 2 Massachusetts Ave, NE
> Washington, DC 20212 USA
> Tel +1.202.691.7523
> FAX +1.202.691.7426
> Email Gillman.Daniel@xxxxxxx
> -----------------------------------------
> "The enemy isn't conservatism.
> The enemy isn't liberalism.
> The enemy is bullshit."
> - Lars-Eric Nelson
> ------------------------------------------
>
>
>
> -----Original Message-----
> From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx
>[mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx] On Behalf Of Ed Barkmeyer
> Sent: Thursday, January 10, 2013 2:32 PM
> To: ontolog-forum@xxxxxxxxxxxxxxxx
> Subject: Re: [ontolog-forum] What is Data? What is a Datum? 2013-01-09-0930
>
>
> On 1/9/2013 8:36 AM, John F Sowa wrote:
>> Sjir,
>>
>>> I would just have said that a datum is a proposition that is taken to
>>> be, or asserted to be, true. The context for that role is any
>>> context in which the proposition is taken to be true.
>> I agree.
>>
>> I would also add that not all data is propositional. For example, the
>> list of names and numbers in a telephone book consists of paired
>> instances of two kinds of data. Each pair becomes a proposition when
>> the instances are inserted in an appropriate schema:
>>
>> "The person named _________ has the telephone number ________."
> I disagree. That is, the meaning of each pair in the telephone directory is
>a proposition of that form, and the pair is a datum. It is not necessary to
>express the sentence per se.
>
> Is John saying that the name of the person is not by itself a proposition? I
>would argue that, if one considers the name of the person alone to be a datum,
>then it expresses a different proposition, to wit: There exists a person
>whose name is X. Further, the presence of the name in the telephone book
>implies the proposition: There exists a telephone number N such that the
>person named X has telephone number N.
> Both of these follow from the proposition that is the meaning of the pair
>(datum).
>
> The distinction I am making is in what the datum is. I argue that a datum is
>a proposition. A value without any interpretation is not a datum. It is a
>child without a meaning.
>
>> And of course, all phone books contain errors at the moment they're
>> printed. Over time, the errors increase until a new phone book is
>> issued. A computerized phone book can be updated more quickly, but it
>> still contains inevitable errors.
> Which only says that not all of the thousands of propositions expressed in
>the phone book are true, even though all of them are apparently asserted.
>
>> More generally, every kind of data is some kind of sign.
> I disagree completely. Every representation of data is some kind of sign,
>yes. But the datum itself is the meaning of the sign. (Thank you, Frege,
>even if Peirce did get there first.)
>
> -Ed
>
> 'I'm sure I didn't mean--' Alice was beginning, but the Red Queen interrupted
>her impatiently. 'That's just what I complain of!
> You SHOULD have meant! What do you suppose is the use of a child without any
>meaning?'
> -- Lewis Carroll, "Through the Looking Glass"
>
> (014)
--
Edward J. Barkmeyer Email: edbark@xxxxxxxx
National Institute of Standards & Technology
Systems Integration Division, Engineering Laboratory
100 Bureau Drive, Stop 8263 Tel: +1 301-975-3528
Gaithersburg, MD 20899-8263 Cel: +1 240-672-5800 (015)
"The opinions expressed above do not reflect consensus of NIST,
and have not been reviewed by any Government authority." (016)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J (017)
|