The Fictions We Create 5: Descriptive sentences and evaluative statements

It is clear that the term *language* has a technical meaning but that it is also used to refer to what humans — but also some other species, e.g. monkeys, — do when they “chatter”! In the first sense, language is a form of communication viewed as acts during which specific messages are passed from one individual to others.

An example often cited in the past was to the “language” of bees, a field of research associated with the work of Karl von Frisch (Nobel prize, 1973) and his discovery that bees inform members of their hive of the location of honey sources by a dance. Here an individual message has content, that is, it is about something, refers to some distinguishable happening or event which may be significant or important to communicants. Yet every message may only be a constituent of a wider language. A piercing cry may be a message — as may be a whimper — but it is not part of a language unless there are additional messages which supplement such singular cries.

A language, in short, usually features many messages and — as we have learnt over the past few hundred years following the publication of several dictionaries — each human language shows that it can grow by leaps and bounds within a relatively short period. In short, a cry can become part of a vocal language, but may not do so consistently. There may be many “cries” which can become an aspect or feature of a language, but need not do so. (For example, “ouch” is a cry of pain, but it is in English.)

A human language has many features by means of which messages are passed from one individual to others of a group, from one to one or among many. Each message has some content — what it is about — but it also possesses a formal structure, which becomes part of its referential meaning. It is therefore incorrect to advertise only the semantic aspect of a term, since each message also occurs in a context which contributes to some degree to what is normally referred to as its meaning for both the sender and the recipient of the message. The prototypical and well-known example of how structure determines meaning is “the cat is on the mat” versus “the mat is on the cat”. There are many versions of this.

As already stated, language is a form of communication viewed as acts of passing messages from one individual to others of a group. But messages by humans are communicated not only by spoken words, but also by gestures, hand signals, etc. Indeed, many modern commentators refer to “body language” to convey that there may be additional meaning to a statement made in written or some other form, e.g., when a poet reads his work in public. Thus one may hear a comment that, “Mr. T asserts ‘xyz’ but his body language tells a different story.” It gives us a different story than if only his words were used to interpret his meaning.

Let us now clarify two terms which may make my position clearer, namely descriptive sentences and evaluative sentences.

Descriptive sentences refer to those elements of a spoken or written language which are presented in the form of statements whose function is to assert that some particular event, x, has a general quality, q. For example, “the sky today is blue” describes x by ascribing one of many possible, suitable, likely attributes to x. Important: the letter q does not refer to an evaluative property of the event (like “beautiful”, “desirable”, “ghastly”, “horrendous”) but refers to a feature which in combination with other features makes the event unique or different from other events, including what it is made of, e.g. a horse of flesh and bones, or a horse made of wood!

Evaluative sentences, on the other hand, refer to sentences of a language which ascribe value to an event, which order or locate that event on a scale of desirability, preferability, usefulness, etc. These are all qualities of the object which reflect its value but does not add to its description. Such evaluations are invariably relative to some other, perhaps comparable, items or events. These therefore do not describe the object by reference to its so-called “defining” properties. To refer to a vase as “in the shape of a bottle with ugly decorations” is a mixed description — and in that sense is no more helpful than talking about it as being “shaped as a bottle with additional decorations”! By adding that the decorations are “ugly”, the object is classified as existing on a scale of values which reflect something about the speaker and not the the object! It is similar to asserting “I love dogs but not cats”, which states a personal preference and clearly has nothing to do with either dogs or cats: in short, being liked by me is not descriptive of either entity or object!

We therefore divide our world into (a) items, events, situation which are describable, can be described, in contrast to (b) rating each of these on a scale which reflects our personal appreciation and reaction to it. We customarily talk about (a) as involving objective description and (b) as involving our personal, subjective, reaction to some event.

In some sense, therefore, descriptive statements may be true or false, whereas evaluative statements only reflect the views and opinions of the speaker, namely, his or her preferences and opinions. To refer to them as being true or false violates the rules of language use. The statements themselves have no epistemic validity, since this quality of a statement cannot be tested without reference to “data”.

Reply to Marc: Facts and Theories

E&E 2
This is my second Elucidation and Exegesis (E&E). In an earlier blog I wrote When we say “This is true” we imply that this statement is defensible, or reliable, therefore should not be questioned. I subsequently received a note from Marc that *defensible,* *reliable* and *unquestioned* do not mean the same. Here is one comment — others may follow.

From Fact to Theory

My earlier statement should be interpreted as saying that if a claim is defensible as well as reliable, it would be counter-productive to question it. What purpose would such a challenge serve?

To demand additional evidence for a claim (1) suggests that the claim — the assertion that such and such is the case — is inadequate and is deemed insufficiently strong, perhaps weak; (2) it also suggests that the case made does not fit snugly into a pre-existing theory. We seem to demand that every new claim falls comfortably into a theoretical place — enlightens but does not disturb our pre-existing vision of the world.

Let us examine each:

(1) Consider first the strength of support for our conclusion “that such and such is the case.” The reason we asked for additional evidence was that the statement went beyond data currently available. One could, in principle, strengthen the data on the basis of which our conclusion was reached by adding to the pile, as it were. But this would be foolish unless we also improved our the method data-collection: a matter of more and better sampling. Data — short for data-points — have to be gathered in sufficient quantity to justify any general conclusions, as for example, “The data can be stated as follows…”. Presenting only a summary of the data assumes that it is the best summary and that a closer examination of the data would not unearth new aspects of events. This assumption has been faulted time and again.

In many cases the strength of support for a conclusion is barely adequate — a discovery often made soon after it was “tabled”. In other cases the method of discovery used may seem “primitive” by current prevailing standards. It happens with alarming frequency, in step with our current technological revolution in the art of measurement. Strong support, on the other hand, leads one to conclude that a claim has sufficient merit — should be considered seriously — may therefore be tagged as factual in the traditional — even, in absolutist — sense. We would then readily concede that whenever a claim is deemed insufficient, adjunctive materials could remedy this and therefore requires to be gotten. If all goes in the anticipated direction — the claim could be moved to the positive side of the ledger. It is all very iffy and tangential. Note: a transition from “perhaps as claimed” to “most likely as claimed” — or some equivalent affirmation — is modal decision — not absolute.

From this perspective stating that “It is a fact that…” — and its milder form, “it could be that…” — reflect different degrees of certainty regardless of whether the method used to establish the conclusions can overcome critical objections. Let us label this our concerns about the adequacy of our investigatory techniques, as when medical researchers are required to run double blind controls to support their view that a drug has the effect claimed for it. Simply to claim that drug x has a lasting effect on the severity of, say, psoriasis is not enough. The claim needs to be underpinned with data gathered in a highly prescribed manner.

Looking over past claims it appears that (a) many, or most, conclusions based on “investigations” remain open to the criticism that the methods used to establish a particular (general) conclusion were inadequate as judged by current standards; (b) that conclusions which appear compatible with an extant theory and which derive their claim from this context operate under a double jeopardy: the one already mentioned (see (a) and the jeopardy that the theory itself may be wrong. To fit snugly into a faulty theory appears to me like catching double pneumonia.

(2) Fitting snugly. A claim, or assertion that such and such is the case has the earmarks of a fact, but it is also a claim that it fits snugly into a pre-existing theory. (Logically, if Tc is a current theory, Tb would be its predecessor. Presumably, the difference between Tb and Tc is noteworthy! If a reason has to be given for changing Tb to Tc it would include that Tc accommodates data which were excluded by Tb and could not be readily accommodated. Indeed — the theory Tc may make up for, or be deliberately used, to handle new data. I will examine this in more detail in a future blog.

Fitting into a Theory vs Finding a Theory to Fit into

If a fact is said to “fit” into a theory this is tantamount to saying that the fact as stated by the proposition p, or set of propositions, sp, is deducible from the premises (foundational base or axioms) of an extant theory. I am careful to distinguish between an extant theory and an ad hoc theory although is not always easy.

An extant theory is one that is in use at the time a particular fact is being accommodated, whereas an ad hoc theory is one that will have to be constructed, or formulated, for the specific purpose of accommodating proposition p into a network. One requires, as a matter of course that the network has some “range,” covers several quite distinct phenomena, otherwise the theory would be solely constructed to accommodate a solitary particular and troubling fact: it is an unacceptable proposal. Even an ad hoc theory is marketed to cover more than one fact, but is so constructed that it aims to bring a particular fact into relation with others. It does so by exclusion: the ad hoc theory usually accommodates some older facts with the new fact, but does so by excluding other facts which appear to be incompatible with the new. It is a matter of compatibility. If a fact is said to “fit” into a theory this is tantamount to saying that the fact as stated, that is, the proposition p, or set of propositions, sp, is deducible from the premises (foundational base, or axioms) of an extant theory.

I draw a distinction between an extant theory and an ad hoc theory by assuming that an extant theory is one that is in use when a particular fact is being accommodated, whereas an ad hoc theory is one that will be constructed, or formulated, for the specific purpose of accommodating p — and related statements — into a network. The ad hoc theory is usually wily — it is deliberately constructed to accommodate some older facts with a new fact, but it does this by excluding those facts which appear to be incompatible with the new. An ad hoc theory invariably creates its own set of facts, and thereby relegates many older facts into the class of “factoids,” that is the class of have-been-facts, a part of “dead science.”

Fitting a finding into a theory is always and inevitably done during the process of publicizing a finding. It is therefore an inevitable consequences of the process of communicating a discovery to others, of stating it, of trying to relate a finding to other matters which are already known, matters which have their establish location in a “knowledge-system” or a “belief system.” One expresses a finding in words and terms which inadvertently position it in a familiar context and in doing so one frames the finding, gives it theoretical relatives.

This description of what happens when a finding is brought to our attention suffers from a major defect: it suggests that a finding is placed into a system, a knowledge or belief-system, like a card placed into filing-drawer, where it has its “proper place.” But — as I shall argue in more detail on another occasion — this is not so: the term “system” is quite misleading and more formal than deserved. It is part of wishful thinking: we like to think of ourselves as “systematic thinkers, as being “well organized,” as knowing where things belong. But this self-image is less true now than ever before: we organize knowledge in many ways, in bits and pieces, as chunks, in conveniently placed separate drawers. A veritable potpourri, but not a system — and we harbour the idea that one day, some day, these different items will all be brought together, like petals of a daisy or cherries on a cherry-tee. No one has succeeded so far, although many have tried: it seems to be a fantasy without end.