[lit-ideas] Statements True False and Otherwise

  • From: Eric Dean <ecdean99@xxxxxxxxxxx>
  • To: <lit-ideas@xxxxxxxxxxxxx>
  • Date: Thu, 13 Dec 2007 16:53:27 +0000

I appreciate Walter O's extended comments on my lengthy discussion of the 
generalization of his erstwhile innocuous assertion about a (particular) 
statement.  I called the generalization S: All statements are either T or F 
irrespective of our ability to establish their truth or falsity.

Perhaps a story would help make my point clearer.  

Many years ago, when Arthur Andersen was still a very widely respected Firm, 
and still deservedly so, with a large and thriving business around the world, I 
was a new partner in the Firm who had just been given responsibility for 
managing the operations of our global computer systems, i.e. the ones we used 
to run our own business (not the ones we built for clients).

Shortly after I got this responsibility, something went wrong with the 
accounting system.  I don't remember exactly what it was, but the net effect 
was that the partners' income for one month was mis-stated.  Since Andersen was 
an accounting firm, it needn't have been that much of a mis-statement for there 
to have been quite a hullabaloo.

Since I'd just gotten the responsibility for overseeing all that, I had to ask 
a lot of people a lot of questions to find out what went wrong.  As I learned, 
there were several systems that were tied together in order to produce the 
financial reports which included the partnership's (and therefore the 
partners') income statement.  Something had fallen through the cracks between a 
couple of those systems, resulting in the mis-statement.

Easy enough to fix once we knew what had happened -- we restored the data that 
should have been passed from one system to the next and re-ran the rest of the 
systems, producing the corrected results -- but one of the questions I had to 
ask, after we fixed the problem, was how come no one had recognized that 
anything had fallen through the cracks until the partners had received their 
income statements and the hullabaloo had commenced.

In the course of those conversations I was asking the guy who ran the data 
center how come our production systems didn't have the controls that would have 
alerted us to the data being lost before we sent the income statements to the 
partners.  He said "Oh, but they do."  I said, with a due measure of 
skepticism, "Then why didn't we stop the income statements from going out?"  He 
said, "Because the system that produces the income statements isn't a 
production system."

That was a surreal moment.  The phrase "production system" was, and still is in 
many organizations, a term of art.  A production system is one for which the IT 
organization has accepted responsibility to ensure that it runs reliably and 
consistently -- it runs when it's supposed to run and when you put the right 
data in you get the right data out.  Since the partnership income statement 
tells the owners of the company, the partners, what their take is, one would 
think that the system producing that statement would be a production system if 
anything was.

The more I asked, the more I learned about the long and painful relationship 
between the Firm's controller and the guys who wrote the accounting system we 
used.  They had been working together for twenty years -- I was the newcomer, 
and several years younger than the combatants.  They were all concerned to 
enlist me on their side of the squabble.

But the simple fact was that whether the system producing the financial 
statements was a production system or not, we all -- the Firm controller and 
the IT group -- were responsible for ensuring that the results it produced were 
correct.  The partners weren't going to care whether the financial statements 
were a production system or not.  They just wanted the results correctly 
reported.  Moreover, even if that system was not a production system it was 
still possible to ensure that it produced accurate results.  "Production" is 
just one technique, or one collection of techniques, for ensuring accuracy, 
it's not the only possible technique.

The reason I tell this story is that my discussion of 'S' was motivated by 
considerations related to the situation in the story.  I think that 'S' works 
when one construes 'statement' as being the sort of thing about which S is true 
-- i.e. by making 'statement' a term of art such that a statement is exactly 
the sort of thing which is either T or F whether we can establish its truth or 
falsity or not.  It is, as it were, a feature of 'production systems'.  

However there are lots of things which appear to be statements but for which S 
is not true, for example "I like vanilla ice cream", to use one of Walter's 
examples.  These are, as it were, the 'non-production systems'.   The problem 
is that for some sentences the fact that they express things standing on one 
side or the other of that boundary is a historical accident, like the fact that 
the financial statement report system was a non-production system.

Perhaps, at a particular point in time, sentences could be divided up, roughly, 
provisionally, into the following categories:

(a) Sentences that express statements (i.e. sentences which conform to S);
(b) Sentences that appear to express statements but which do not in fact 
express statements in the sense of class (a) (e.g. "Walter likes vanilla ice 
(c) Sentences that appear to fall into category (b) but which either now or 
shortly will fall into class (a) (perhaps because we learn something new about 
our body chemistry, etc.)
(d) Everything else (questions, commands, etc.)

It's category (c) that I was concerned about when I was asserting that S does 
not apply to all statements.  If we human beings could know that we could today 
define class (a) for all time, why then class (c) would be empty and if (c) 
were empty, then my concerns would be moot -- there would be no sentences that 
appear (on the surface) to express statements but also appear (at the next 
level of analysis) to express something that's not a statement but, upon yet 
further understanding, turn out to express statements after all.  

But we have a planet full of corporations who are trying their damnedest to 
turn a lot of (c)'s into (a)'s -- a lot of market research amounts to just 
that.  I think that a lot of nonsense, and pernicious nonsense at that, gets 
justified by people demanding that class (c) must be large, though I also think 
that class (c) is not empty (otherwise we'd know at least a lot more about (a) 
than I think we do).

Corporations and their boards of directors who, lest we try to distance 
ourselves from them, represent the shareholders which includes many of us on 
this list by way of pension funds, 401k's, mutual funds, etc., demand that 
everything must be 'objective', by which they mean expressible in sentences 
from class (a), and 'measurable' by which they mean amenable to being 
engineered, all of which is just an elaborately disguised form of trying to 
manage the whims of the goddess Fortuna...  

The problem is that most of what precipitates poor outcomes in companies 
involves things expressed by sentences from classes (b) or (c).  Note, 
moreover, that just because a sentence may be in class (b) doesn't mean that a 
particular token of it might not actually be in class (a) -- i.e. on some 
occasion, Walter may be making a statement with "I like vanilla ice cream," 
which statement might be translated into something like "at the moment, all 
other things being equal I will choose vanilla over other flavors of ice cream" 
or something like that.  

Anyway, the result is that if one really does want to manage successfully, one 
needs to deal directly with the reality one is confronted with, irrespective of 
whether that reality is best expressed by sentences of type (a), (b) or (c), 
but what 'dealing with' means varies by whether the relevant reality is 
expressed by sentences of type (a), (b) or (c).

In other words, whether the system is production or not, the results still 
matter.  My no-doubt obtuse way of making that point was to say that just 
because a system isn't production doesn't mean it's not a system -- i.e. just 
because a statement doesn't conform to S doesn't mean it's not a statement.    

Regards to one and all,
Eric Dean
Washington DC

Other related posts: