11. WOODHOOD CUBED?

(Part of a series based upon Stiles, The Anatomy of Medical Terminology (Radix Antiqua 2015; ISBN 978-1-988941-240)

             Our discussion of "wordhood" (see Posts 3 & 5) leads us to yet another consideration.  Now that we are "scaling it up," what if we move away entirely from the concept of generic words like "bone" or "tissue" which name general categories with many specific members, and consider what we might do with what could be called a generically describeable set of words exemplified by at least one well-attested specific member?  For example, any Medical Dictionary will show us (by including it) that the word

            peri-oste-um               a part surrounding a bone

is a so-called "real" word.  Now, the "generic" features of this construction are two:

            1) the termination -um, denoting a physical object, here in conjunction with osteo- naming a part somehow associated with a bone (either belonging to it, or in some other way involving it);

            2) the prepositional prefix, which by definition restricts the nature of the "association" to "locational involvement."

            It might seem that a new kind of rule can be hypothesized here, this time having at its smallest scale nothing to do with specific members of a generic category like the one called "bones," but rather with specific body-parts (or other objects) named generically by the addition of any specific PREFIX denoting LOCATION to the "word-remnant" -osteum:

                        If the set of real words in Medical Terminology includes one word of the form

                        P-oste-um (where "P" represents a PREFIX denoting LOCATION),

                        then that set also includes

                        all words of the form P-oste-um.

            Unfortunately, an actual count "in the Dictionary" reveals that only two of our postulated twenty or more words have so far been attested.  Eighteen words "missing" out of a potential set of twenty is pretty dismal arithmetic, and would seem to augur badly for our rule. To try to salvage it, we might begin by replacing "all words of the form P-oste-um" by the phrase "a large number of words of the form P-oste-um," on the grounds that by getting scrappy we might be able to convince you that some number of such words are "real" even though unattested.  But this touch-up would only return us to the (possibly circular) arguments of the "fighting words" post; therefore, now that we've seen something about the effects of change of scale, let's also try moving toward a "meta-rule" from these admittedly discouraging beginnings.

            The rule we can postulate by as it were "zooming out" from P-oste-um is of course formed by replacing -oste-um by something more general.  For -oste- we can read  -X- (designating as always "any organ or part"); and since the termination -um is only one of a number of nominative singular endings -N which all function the same way in naming real, concrete, objects, we can write our new meta-rule formulaically:

                        If the set of real words in Medical Terminology contains even one word

                        of the form P-X-N (where -N stands for any nominative singular ending),

                        then that set also includes

                        a large number of words of the form P-X-N.

            Notice that in this rule we are no longer making all-inclusive predictions: "a large number" is probably as specific as we can be.  And to prove to you that what I'm trying here is not a kind of verbal sleight-of-hand, I want you to remember that our first specific example of this very meta-rule generated a measly 2 out of a possible 20 or more predicted words--I am not trying to baffle you into forgetting this!

            Now let's make the additional observation that, at "scale," all of the words generated by our new meta-rule mean "the same thing;" that is, any word

                        P-X-N (given the definitions previously described) will name

                        a body part or other physical object involving X in the locative way

                        specified by P.

Thus, for example,

            end-oste-um                a part inside a bone

            peri-enter-on              a part surrounding the intestine

            para-cortex                 a part beside the cortex

            Note that on the dimension of translating from medical terminology into English there is nothing new here, nothing that isn't predicted in Chapter 13 of The Anatomy of Medical Terminology; what is new here is that we are using the as it were "generic" translation as the base for a "meta-rule" to generate medical words.

            Confused?  To "touch bases" with reality for a moment, a quick examination of the actually attested relevant words ("in The Dictionary") associated with just three of our locational prefixes (endo-, peri- and para-) gives us fifty words which fit our pattern (specifics available on request).  While this number is arguably "respectable," it is still discouragingly small given the huge scope of our category -X-.

            But remember that our overall thesis here is that the set of all real words is far larger (orders of magnitude larger) than any set which will ever be found in any dictionary.  At the particular scale we are discussing here (P-X-N), we can begin with a simple (and non-controversial) observation: if there is an adjective formed by the usual Latglish method of adding an adjectival ending (like -ic or -al) to a combing form -X- denoting as always an organ or body part, then there must be a corresponding noun formed by adding a nominative singular ending (whether regular or irregular) to that same combining form.  Thus, for example

            crani-al                        presupposes                cranium, and

            cortic-al,                                                          cortex.

Though not all of these words have been attested in English (never forget that Medical terminology is a kind of English), they do exist in Latin; and I hope it can be accepted for this argument that any Latin word naming a body part is also a real word in Medical Terminology--a word "waiting to happen," perhaps.  A generation or two ago, for example, cost-a "rib" (answering to  well-attested cost-al), and clavicul-a "collar-bone" (or "clavicle;" presupposed by clavicul-ar) were not "in the dictionary;" but they are now (although my impression is, they are still not in very general usage, in comparison to words like skelet-on or thorax).

            If this observation is accepted, then we can postulate that something analogous is the case with the "locative" subset of those adjectives, the subset which is characterized by an introductory prepositional prefix P-, and the meaning "located in the prepositional relationship to X specified by P."  Here are some examples (all "in the dictionary"):

            para-ven-ous               beside a vein

            peri-ven-ous                surrounding a vein

            endo-ven-ous              inside a vein

            The word-pattern exemplified here can be written P-X-A (where -A denotes an adjectival ending).  A further proposition arises: where X names a body part, then in the real world of concrete objects (regardless of nomenclature) there will exist a body part (or other object) describable by every single adjective of the type P-X-A.  For example, we could refer to perivenous tissue, or an endovenous blockage, or a paravenous swelling.

            To put this another way--and here comes the "gotcha!" moment--if the set of real things in the real world contains all organs or body parts named -X-, then that set also contains all real parts or objects P-X-N, that is, "a body part or object located in the prepositional relationship to X specified by P-.

            Note that we are not claiming here that the particular words generated by our three examples are actually attested (in fact, they are not, despite their transparency):

            para-ven-a                  something beside a vein

            peri-ven-a                   something surrounding a vein

            endo-ven-a                  something inside a vein

--we are only claiming that the concepts so named are well-documented; and we note in passing that the Latin noun ven-a is among those (like cost-a and patell-a, discussed above) which have been or could be used in English as well.

            At this point, let's do some actual counting, restricting ourselves for this experiment to the same three prefixes as before.  Of the fifty words we found earlier, of the pattern P-X-N, ten could be said to be "predicted" by analogous P-X-A; but there are about 225 words altogether of the shape P-X-A.  In total then, the words counted here can be said to specify 265 body parts or other objects, and to at least strongly imply their names!

            If this is confusing, let's try a slightly different approach.  Other noun-endings are of course found on words beginning P-X-, terminations denoting abstract nouns naming processes, procedures and the like.  For illustrative purposes, I counted the occurence (in our same three-prefix sample) of just one of those, those formed by the addition of the abstract noun termination -itis, namely

            P-X-ITIS          the inflammation of a body part

                                    located in the prepositional relation to X specified by P-

                                    (or, the inflammation of something involving an object

                                    located in the prepositional relation to X specified by P-);

this count (omitting words of this form possessing previously counted analogues shaped P-X-A or P-X-N) amounted to more than 50 more words.  Again, such words "predict" or presuppose a concrete noun, as in the following sequence:

            P-X-ITIS            = the inflammation of (something involving) P-X-N,

            where P-X-N    = something located in the prepositional relation to X specifed by P-;

            therefore, if P-X-ITIS is a real word, so too is P-X-N.

            To sum up, I have tried to make the case here that the existence of even one word of the form P-X-N presupposes the existence of many more of the same form; supporting evidence can be marshalled by considering that for every analogous word (in our discussion, P-X-A and P-X-ITIS) there exists a word P-X-N, and without even looking very hard we've found nearly 300 examples that would be difficult to defrock as "real words."

            But what about words of the shape P-OSTE-UM?  O yeah; next post!

                                                                        - o -