(From Lognet 94/3. Used with the permission of The Loglan Institute, Inc.)
Letters policy: Unless otherwise stated, letters addressed to logli in general, The Institute, or any editor of Lognet will be considered as offered for publication. But it would be good if the writer explicitly offers. We reserve the right to edit letters, mostly just to drop material that has to do with ordering books, etc. Sometimes a given correspondent will have several letters in the hopper, so to speak, and we will combine them into one purely for the sake of clarity. If you are not on e-mail and your letter is a long one, we’d be grateful if you’d enclose a soft copy of it on a disk. We can translate most word-processors and having your letters on disk could save us a lot of typing.
The first letter is a composite one from Bob McIvor, our Takyrultua, and is a continuation of three threads of discourse—(1) on the uses of illogic in nature, (2) on the human language biogram, and (3) on the importance of face-to-face interaction in language learning—all originating in earlier issues.
Hoi Djim, fie,
With regard to the human language biogram, I recall reading, I believe in Scientific American a couple of years ago, that studies indicated that verbs and nouns are dealt with in different regions of the brain. If we are wired this way, it is not surprising that a language which treats these elements as indistinguishable would have a hard fight getting established, since the built-in program would need to be overwritten. However, I suppose these studies were done with Indo-European speakers. I recall wondering at the time if studies with speakers of languages where the noun is often built into the verb (much like the idea of “shoulders” is built into the idea of “shrug”, which doesn’t stop people from adding “shoulders” anyway; can one shrug anything else?) would have given different results. I believe Hopi and Navajo are such languages, but this is hearsay and not personal knowledge. [We have a Hopi specialist, Dave Seaman, among our members. It would be good to hear from you, Dave, on this interesting point.]
... A further example of how “illogic”, in particular, PHEPH and affirming the consequent, aids survival: Bird eats Monarch butterfly. Bird gets sick. PHEPH: Bird “believes” sickness caused by eating butterfly. (Happens to be true, but not good logic!) Affirm the consequent: bird infers that all butterflies of this pattern are sick-making. (Happens to be false; Viceroy butterflies, which also display the Monarch pattern, are not sick-making.) Bird avoids eating all butterflies with the Monarch pattern. Result: bird is not ill, even though assumption is false. Side effect: Viceroy butterflies, by protective mimicry, avoid being eaten.
... I neglected to comment on your suggestion some time ago that we need more face-to-face teaching, preferably a “colony” of speakers. Indeed we do need to develop daily speech in L, but, aside from students like Mike Demoulin who can apparently still round up study groups, I don’t know how this is going to be achieved. In the meantime, if we can encourage the production of more L texts, both original and translations, these have in the past shown up constructions which were difficult to say in L, and so have often led to useful grammatical developments.
We had just four “takers” of the colony idea ( LN 93/3:4ff), three quite tentative, of whom two—Bob himself and Wes—offered to be occasional visitors. The two interested in actual residence, and one of these a bit dubiously, were Alan Gaynor (who, remember, wrote an LN article on the thoughts provoked in him by the colony idea; see LN 93/4:4ff) and Gilles Dignard, our French-Canadian logli, who was the one potential colonist whose only condition was that there be others. Uu, there weren’t any others! But thank you, Gilles, for being our first declared potential colonist! So crucial though it might be to Loglan’s development as a living language, a “logle”—an actual Loglan district where logli live and L is spoken—is apparently not to be yet.
The next letter is from Michael Demoulin, the logli who has actually formed, and is now leading, the one Loglan study group in existence at the present time: the one at USL (which I think means “University of Southwestern Louisiana”; but I can’t confirm this, as I seem to have mislaid my universal codebook, soi crano).
Hoi Djimbraon, fie:
I’m afraid that the [USL] study group has stagnated this semester. This has been my busiest semester, and another member had to take a semester out to get his job going, and so forth.
We do plan to restart again next semester. The group had been a group for talking about Loglan, not in Loglan. I think that a renewed commitment on our parts should get the group where it should be.
There are three prospective new members for the group so far. One thing that will make the group go a whole lot easier is installing LOD and the MacTeach programs on our UNIX system. Are the programs available for UNIX? If not, I can convert. If we can get the UNIX-ware, would we have to pay again? (That might be a good idea anyway, because of the new members.)
Isui, how are the plans for La Logli going?
No; you wouldn’t have to pay again! Yours is probably the most vital project currently going on in Loglandia, Maik, and we here at The Institute—all of us volunteer workers (whose Loglan interactions are unfortunately confined to e-mail these days)—wish to encourage you face-to-facers, and any others wishing to form such living study groups. (Are there any others in the offing? Please report in and let us help you get started.) Now I’ll turn your technical questions over to Dr. McIvor.—JCB
Re running LOD, LIP, and M1-3 on UNIX, the answer is that the DOS versions of all these programs are in ANSI-C, and do not, as far as I know, use any DOS-specific code, so should require a minimum of effort to run on UNIX. Randall was given the source for LIP, at least, with the intention of making it UNIX-compatible, but I don’t know if he has got around to it yet. Write him at the e-address given on the back cover to find out. If you need the source code for any or all of these programs, and you already have a “non-disclosure” agreement on file with The Institute, just e-write me and tell me where to send the code. If you haven’t yet filed such an agreement, you may e-write our General Counsel, Wes Parsons, for one (see back cover); and when that is done, I’ll be pleased to send you the code. —RAM
The following paragraph is part of a letter from Emerson Mitchell, a lodtua (logic-worker = logician) and a new member of the Lodgru (logic-group), that subset of our tiftua (offering-workers) who are now hard at work discussing various new logical usages. Emerson sent this letter to the Lodgru last November, and I thought this paragraph in it was such a useful clarification of the relationships among grammar (especially lexemic structure), metaphysics, ontology, and the Whorf hypothesis that I asked Emerson if I could reprint it in Lo Lerci; and he said yes.—JCB
... Please do not tell me that Loglan has no assumptions, any language does. Any language assumes, at a minimum, that different words in the same syntactical class bear similar relationships to other [syntactical] classes. For example, putting past and future both into the “tenses” category assumes that past and future have similar effects in the grammer [and on the world]. This is what Whorf noticed. His hypothesis is that these assumptions built into the syntax and semantics of a language have measurable enabling or restricting effects on the behavior and thinking of people using the language. ...
Actually no one had said that Loglan “makes no assumptions”, but the possibility was evidently in the air. In any case, this sort of preemptive clarification is always worth tendering at a roundtable of this kind. It’s just the sort of thing we try to do for each other in the free and easy atmosphere in which the Lodgru does its work. Would anyone else like to join these roundtable discussions on the best ways of compressing logic into speakable Loglan?
The next letter is a long one from Paul Doudna and was written in response to my response to his letter in LN 94/2, and it now has a new set of my responses (responses to responses to responses!) inserted. Our interchange takes up some of the same matters broached last time; but this time Ex-Member Doudna and I seem to have arrived at some interesting conclusions and rapprochements. It shows that if you keep at this communication business long enough, soi crano, good things can happen.
I was pleasantly surprised to receive a copy of Lognet 94/2 and to find my year-old letter ... published with a response. I was also disappointed to find the response did not really address the issues I had originally [raised] in 1989 after reading Loglan 1. Of course the discussion of “binary metaphor” was informative and gives some further insight into the philosophy of Loglan. Here are my responses to your response: [Passages quoted by PD from my LN 94/2 answer to his original letter are indented in plaintext as separate blocks, while my current responses are italicized and inserted (as this one is) in the text of his letter.—JCB]
The writer had purchased the 4th Edition of L1 in 1989 but on reading it, had apparently decided not to (re)join The Institute. As a consequence he received no further messages from us and we apparently seemed to him to have slumped into total inactivity!
There were a number of reasons why I didn’t rejoin The Institute. No response to questions about Loglan; no acknowledgment of certain inconsistencies, unclear explanations, and typos in Loglan 1 that as an author you should want to know about; ... an overly secretive and exclusivist attitude on the part of The Institute; an endlessly changing language that never seemed to stabilize.[Thank you for telling us, soi srisu. It’s good for us to know what reasons some members have had for not renewing. Probably when your letter arrived in ‘89 it was both too long and too negative to inspire immediate attention, and so got filed away...no doubt with the best of intentions, uu, for an “eventual answer”. I hope this counts as one, soi crano.]
I’m glad to see The Loglan Institute is still active. I only hope that others won’t be made to feel like Rodney Dangerfield. (I don’t get no respect!)
It is true that we do not have the kind of public relations resources that, say, the Ford Corporation has, soi crano, that might have kept even non-members apprised of our existence.
You don’t need to act like the Ford Corporation; just use common sense. Some things turn people on; some things turn people off. If you want to promote Loglan, and I’m sure you do, change your PR. [Thanks for the advice, soi drani crano, but there’s a limit to what our resources permit us to do in the way of projecting a public image, and we had to confront it long ago. We decided then that as part of our GPA (“Going Public Again”) those of us who could would write articles about Loglan. Reed Riner’s Fall 1990 Et Cetera article“Loglan and the Option of Clarity” was written in that spirit; and Greenhood and I mention L in our 1991 article about the evolution of language (JSBS 14(3):255-309). One or two other articles with possible PR value have been written and are still coming up. One is about to be published in La Logli...which is preaching to the converted, I know; but still, it might be republished elsewhere. The rest of our PR is apparently word-of-mouth. Still, we must be doing something right. We’ve been growing in a very steady way ever since our GPA in 1989, and this means that old members are renewing as well as new ones coming aboard. In any case, our membership now is larger than it’s ever been, and we’ve got a larger and more dedicated crew of tiftua (volunteers) than we’ve ever had before. We think one good reason for both these happy results is that this is a pretty good journal, soi no selsai kubcra. By the way, we’d love to have you aboard again, Mr. Doudna. You evidently have much to contribute.]
“Soi crano.” Whenever I see this I keep asking myself, “I wonder what he means by that?” Does this have logical significance? Is it an inside joke? A shiboleth perhaps? Is it a meaningless expletive like, that is, ewe no? [The expression soi crano means I’m smiling as I’m writing this and soi drani crano means I’m smiling dryly, as those expressions might be inserted parenthetically in a piece of English text to advise the reader that the writer is smiling or smiling dryly as he or she is writing, that therefore he’s not quite serious and may even be joking. The operator soi takes as its operand any predicate expression (such as no selsai kubcra (immodestly wide-smiling (grinning)) as used above). When used honestly, that predicate expression will describe one or more things that the writer is actually doing or being while writing—thus soi srisu (I’m serious) might be used when one actually is serious but might be suspected of not being—and will thus add a rich attitudinal dimension to the ordinarily narrow bandwidth of “visible speech”. Soi and its inverse sue were proposed by Steve Rice in 1990, and were immediately and gleefully adopted by the Keugru; see the Sau La Keugru column in LN 90/1. Back copies of most issues of LNs are still available.]
This famous problem [concerning the ambiguity of ‘pretty little girls’ school’] ... had been “much discussed” in 1976-77, and, we thought, resolved.
The fact that you (and I suppose most others in the inner circle) thought this problem was resolved reveals a rather fundamental difference in the way people view linguistic analysis. [The logli who “thought this problem resolved” were not an inner circle but the readers and contributors to The Loglanist during the years 1976-78, when PLGS was being actively and publicly discussed. All the solutions proposed during those years were published in TL, and were there discussed quite openly. One of these was finally accepted as an addition to the language in the informal, pre-Keugru way we had in those days...namely by TL’s editor John Parks-Clifford and I tacitly agreeing—usually at the end of some long public discussion—that we had arrived at a satisfactory solution to some problem, and then “adopting” it by including it in the next “official description” of the language. That took place in the case of the ge/ci/cue solution to PLGS in TL4/3:32-34 in November 1980, in case you’d care to have a look at it.] The phrase PLGS was apparently viewed as being merely a generic string of four terms having some unspecified binary structure. [That’s not correct. The unmarked string bilti cmalo nirli ckela has never been viewed in this way. For how it is viewed, let me refer you to the cited 1980 document...or, for that matter, to L1 itself, 4th Edition, pp. 129-34. The string-marking system described in these places was altered in a minor way in 1993 when geu replaced cue as the righthand marker of “internal groups” in strings.] The particular meanings and functions of the four terms were regarded as irrelevant. The basic problem was to find some neat and pronounceable way of mechanically representing the five basic surface structures: P(L(GS)), (PL)(GS), ((PL)G)S, (P(LG))S, P((LG)S). [These are not “surface structures” in the Chomskyian sense I assume you mean, but “deep structures”.] Having done that, and in the process producing 12 or 17 additional (but in my opinion, somewhat dubious) interpretations, the problem was considered “resolved”.
I attempted in 1989—totally unsuccessfully apparently—to explain why the PLGS problem was not resolved. But let me try again. Consider the string “interesting little red school.” It is not ambiguous. [But it is, ue!] We are speaking of one object X, having four attributes: X is a school, X is red, X is little, X is interesting. [If this were what the E string X is an interesting little red school meant—which few E-speakers would say it did—then it could be clearly rendered into L, not by modification but by logical connection: Da treci ce cmalo ce redro ce ckela = X is interesting and small and red and a school; and this has the deep structure—as LIP would show you—of da (((treci ce cmalo) ce redro) ce ckela simply because ce forms “afterthought” connections, i.e., is left-grouping.] The surface structure of the string is I(L(RS)). [That’s the structure of one possible meaning, but it is not the structure of the one you think ILRS “unambiguously” means, which is left-grouping.] The other four possibilities of binary grouping make no sense. [But some of them do, ue!]
On the other hand, each of the five PLGS structures involves two objects, X and Y. One attribute applies to X (X is a school) and one attribute applies to Y (Y is a girl). The other two attributes may apply to either X or Y. We don’t know which without indicating the structure; hence the string PLGS is ambiguous. A further complication, of course, is “pretty”, which may be either an attribute of an object or an attribute of another attribute. The relationship between X and Y is the somewhat vague “possessive relation”. P(L(GS)) and I(L(RS)) have two distinct structures. [What makes them distinct is that different types of “modification” are indeed involved; but this is a semantic matter—and could ultimately be a lexical one (when and if we adopt a set of “metaphor-typing infixes”, which I’ll mention in a moment); but their grammar would even then be the same.] The first contains four attributes, two objects, and a relationship between the two objects (specified in English but not in Loglan); the second contains four attributes and one object. Common sense or nonsense? You decide.
This letter [in response to receiving a free copy of Lognet] is one of the many interesting results of that campaign to reawaken sleeping logli.
Reader Doudna [says Reader Doudna] is, I sup-pose, a “sleeping logli”. He doesn’t mind waking up if something interesting passes by. “As I understand it, Loglan has eliminated syntactical ambiguity from the structure of the language. That is, one can distinguish A(BC) from (AB)C.” That’s a quote by [JCB] from my letter. Well almost, anyway. What I said was: “... Loglan has eliminated syntactical ambiguity from the surface structure of the language.” If Writer Brown doesn’t believe that Loglan has a deep structure, then the word “surface” would be either meaningless or a redundancy. But the misquote rather changes the point I was trying to make. [I am beginning to see that the distinction Reader D is making between “deep” and “surface” structures is the one that other logli treat as the difference between “lexical semantics” and “grammatical semantics”: that is, between what is conveyed by the choice of a word to fill a “slot” in some grammatical structure, and what is conveyed by the choice of that structure. Thus, DA PREDA is a grammatical structure. Choosing it in the course of speech or writing relates two lexemes (word-classes) DA and PREDA in a certain way, i.e., as the “first argument” and “predicate expression”, respectively, of a “predication sentence”; this is a grammatical choice, presumably related to the speaker’s intention to claim something about something. The choice of ta and mrenu, however, as occupants of the DA and PREDA slots, respectively, in that chosen structure are lexical choices made by the speaker to fit his particular use of it; and the two sets of choices together produce the claim Ta mrenu = That (person) is a man. “Deep” vs. “surface” structure is not involved here...as, in fact, it almost never is in L. All surface structures in Loglan reflect Chomskyan deep structures and vice versa. That is what is meant by saying that the logical structure of L speech “lies on its surface”, that we know what we are claiming simply by listening to what we say. This identity of deep and surface structure is probably L’s largest gift to clear thought.]
Reader Doudna has a legitimate complaint here: we should be more careful to specify the kind of ambiguity we are talking about when we say that Loglan isn’t.
That’s the first time I’ve heard that admission! Yes, we all need to be more careful in our use of words.
[Syntactic ambiguity] is not only massively present in English ... but it probably also characterizes every other natural language that has an “ancient” grammar, which is all of them.
Careful writers, and readers who are able to make common-sense distinctions between the plausible and the implausible, make syntactic ambiguity a relatively rare problem in English. The serious problems arise when English is used in a highly specialized context, especially with computers. Computers have a hard time with plausibility. [Maybe “serious problems” also arise between (1) politicians and their constituents, (2) spouses, (3) readers and writers, (4) students and teachers,(5) men-in-the-street, and (6) the inventors and users of “logical languages”, soi crano..]
As to semantic ambiguity, I believe Reader D[oudna] is wrong to suggest that we ought to try to banish that sort [of ambiguity] as well. I’m afraid that to do that—and formally, it could easily be done—would be to turn L[oglan] into a sterile logical code, a deductive system.
Reader Doudna [says he] would have been wrong if he had in fact suggested that. But he did not. As to the implication that banishing semantic ambiguity “could easily be done”, I don’t know what to say. It’s a problem lots of AI people are working on and I don’t think that computers would object to a sterile logical code. [The trouble with “sterile logical codes” is not that computers don’t, or wouldn’t, like them but that, while they are in fact easy for mathematicians and logicians to devise, they have no clear relationship to the phenomena of either human thought or human language. Recent writers on “fuzzy logic” have made that point abundantly clear. In short, even the best logical codes given to computers are not good enough to permit them to make the kinds of “fuzzy” but immensely useful judgements that humans regularly make. Loglan goes a step beyond that: L is grammatically logical but semantically as rich and fuzzy as any human language is, and as every human language evidently needs to be. Moreover, it gives us, as Loglan lexicographers, the opportunity, through the multi-place structures of the predicate calculus, to make the meanings of even fuzzy claims clear.] I had always assumed that the reason that semantic ambiguity exists in Loglan is that it is impossible to totally remove it. [Nope.] By comparison, removing surface structure ambiguity is relatively simple. [Right; and it’s been done; this was the “MacGram Project” of 1978-82.] No semantic analysis is required; it’s simply a matter of logical manipulation of symbols.
Metaphor-making appears to be such a powerful and widely distributed human faculty that I do not think that any program to rid spoken language of the sort of ambiguity that metaphors ineluctably generate ... could succeed with humans.
I quite agree about the fundamental role of metaphor in language creation and evolution. ...
If B[rown] & G[reenhood] are right [in their scenario for the evolution of mind and language (JSBS, 1991, op. cit.)], to abandon metaphor altogether would be to abandon language altogether. I trust that even Reader Doudna would agree, soi crano, that such a move would be draconian.
No, Reader Doudna [says he] wouldn’t want to abandon metaphor...or language. It would be worse than draconian; it would be a downright nuisance. “Soi crano.” There’s that phrase again. Actually it’s kind of fun trying to figure out what meaning will fit all of the contexts in which this ubiquitous phrase occurs. I hope, ewe no, it’s not an obscenity.
I would like also to suggest that Reader D[oudna]’s evident hope for a less semantically ambiguous language is quite a reasonable one.
That’s much closer to any potential suggestion I might make...and a lot different from “banishment”. (See previous comment.) Yes, Reader Doudna is usually quite reasonable.
The point of all this is that the Loglan word-building system is quite a different one from that of English. ... The Loglan system for making new words is, as we all know, to invent a binary metaphor—not a unary one—for the required new notion.
Well, I don’t know...that’s hard to believe. English uses a variety of word-building techniques. If you don’t already know about it, you would be interested in the book Fifty Years Among the New Words — A Dictionary of Neologisms, 1941-1991, edited by John Algeo (Cambridge University Press). The introduction contains a very thorough analysis [of] the way words are formed in English. The basic sources are: creating (<.5%), borrowing (2%), combining (68%), shortening (8%), blending (5%), shifting (17%). “Combining” includes what you would call “binary metaphor”, if I understand your use of that term. “Shifting” includes “semantic shifts” or “unary metaphor”. If you cannot find the book I will send you ... a copy (of the introduction). [Thanks for the reference; it looks useful for people in our business. I’d very much like to see a copy of its introduction.]
We seem, in short, to have struck on something fairly useful ... in the way of ambiguity control.
I hope it works, but I am skeptical based on what has happened to binary metaphor in English. One would be very misled to think that a “manufactured article” was hand-made. It is always interesting to learn the roots or affixes that make up such words as influence or hierarchy but that doesn’t necessarily tell us how the word is currently being used. Why should Loglan, assuming it comes to be widely used, be less subject to semantic shift than English? [Because L writers and word-makers—not by edict or explicit agreement, but simply because it is possible to do so—seem to be avoiding semantic shifts (unary metaphors) in favor of building new CPXs on the basis of n-ary metaphors whenever new uses of old, related concepts arise...as they very frequently do in any language. Thus we logli have two words, stutaa and stucue for the two senses of tell, whereas in E, by means of semantic shifting, we try to get by with one.]
Thus I’ve called for some time now for some ambitious logli to build us a set of infixes—to be optionally inserted by the speaker between the terms of some metaphor that [the] s[peaker] has just created—that would tell the hearer just what kind of metaphor [the] s[peaker] was intending.
You are apparently now talking about forming strings by compounding predicate words...what I referred to as a “modification pair” in my letter. I assume that the term “binary metaphor” applies both to these strings of words as well as to the compound words built from affixes. [By convention, a “defining metaphor” in L is a string of words, say terla bidje Earth-edge, from which a CPX predicate, in this case telbie, has been made. But telbie is not a metaphor although terla bidje is; it is the L word for horizon and means X is the horizon of planet/astronomical body Y from position Z on or above its surface. With this sort of apparatus available to logli word-makers there is evidently little occasion for semantic shifting in L.] I doubt if you will find an “ambitious logli” for this job. You are the logical person to build a set of infixes since you know Logan better than anyone else. [That’s not true any more, ui! Anyway, it's not a question of knowing Loglan, it’s a question of knowing, or knowing about, a large number of the world’s languages, which I don’t.] I am not sure this would really fix (pun intended) the problem, but it would help. One approach would be to use juxtaposition when forming one-object binary metaphors and some little word (or better, one from a list of little words) for forming two-object binary metaphors. The simple distinction between one- and two-object metaphors would eliminate a large class of ambiguous binary metaphors.
For example, “brick house” is ambiguous whereas “little house” is not. [Actually, both are ambiguous...if only because all metaphors are.] The first is a two-object metaphor, the second is a one-object metaphor. [You are now approaching what I’ve meant by a “typology of metaphor”. There is a finite set of metaphorical types, I believe, which are used by humans the world over to extend their local lexicons. One of these is the “material + made-thing” metaphor, of which brick house and woolen coat are E instances. These are quite different from the milk bottle-type of metaphor, which we might call the “content + container” type. A third type is the very common “domain-limiting” type, of which your example little house is an instance. Thus “a little house” is not really something that is both little-in-general and a house; it is something that is “little for a house” (and, of course, also a house). This is the “(limited) attribute + (limiting) domain” type of metaphorical image. I could go on; we all could. There are tigru janto, for example, who are neither “made of tigers”, “containers of tigers”, or “tigerish for a hunter”; but something else. You decide what! There are at least a dozen other ways of constructing metaphorical images in English, and probably even a richer set of ways in which metaphors are constructed in metaphor-rich languages like Chinese. These need discovering. I don’t know the complete list. Do you? And after that is done for a wide set of languages, a set of CVV-form infixes needs to be devised and assigned. That's the easy part. Any takers for this lovely, linguistical project? Mr. D, for example?] That is, “brick house” implies [that] there are two objects, X and Y, such that X is a house, Y is a brick, and X is related to Y by some unspecified relation. “Little house”, on the other hand, implies there is one object X which is a house, which is little. If the one-object and two-object patterns are not distinguished, then all binary metaphors are artificially being forced to be interpreted ambiguously since there is no way to know if one object or two objects are intended. English is obviously ambiguous in certain ways in which Loglan is not. But in at least this one respect, English is less ambiguous than Loglan. In English one can generally distinguish the one-object (adjective + noun) from the two-object (noun + noun) metaphors. [This distinction, though interesting and important, simply does not go far enough toward generating a theory of metaphorical types.] “Baby doctor” is an interesting exception. [It’s probably of the same type as tigru janto, which we might call the “served + server” or the “target + shooter” type.] The most likely meaning is a two-object metaphor. A rather unexpected meaning is a one-object metaphor. “Baby doctor” and the Loglan counterpart cindu kicmu (1975 dictionary) are both ambiguous and for the same reason. [(Human) baby doctor would now be humni kicmu, or simply cinta kicmu leaving the humanity of the doctored infant implied.] The surface structure [let’s read ‘grammar’ here] of both English and Loglan are identical in this expression.
I think of these infixes as being strictly optional, of course...that is, I would expect them to be replaced by their null allolex [that is, left out,] in normally rapid speech.
The question is, are binary metaphors (two concatenated predicate words) to be regarded as primary structures, which may be optionally patched up to make them logically correct? [“Patching up” is probably not as accurate a description of what we would be doing here as “typing” or “specifying” what are in nature unspecfied, namely our metaphors...and then, probably, only when we were asked to do so.] Or should logically correct descriptions be regarded as the primary structures, which may be optionally abbreviated when the context will allow it without confusion?
The [phrase] “null allolex” reminds me of the cherry tree which Washington replaced with a null allotree. I am not sure whether the morpheme “lex” refers to “form class” (loosely, part of speech) or to “lexeme” (vocabulary item, loosely “word”). [In our L studies, we use these words in exactly the opposite senses from those you suggest. That is, a “lexeme”, in a sense that has been established in our literature for about 30 years, is a set of grammatically interchangeable words; that is, a set of words whose members can be exchanged for one another without altering the grammatical structure of any utterance in which any of them may appear. Thus mi tu da ta toi toa can all replace one another anywhere without altering the grammatical structure of any Loglan utterance. Another way of saying this is that they are all allolexes of the same lexeme, namely the lexeme which we arbitrarily call “DA”. That’s how we use the word lexeme. On the other hand, the word lex, in this same technical literature, means an allolex of some lexeme, that is, a member of some clearly defined grammatical class of words in some language. Thus a lex is simply a word in some language, and therefore a lexemically classifiable element of that language.]
In any case, if you recognize the existence of the “null allolex”, then it can be replaced with a hypothetical non-null lexeme (or morpheme), say “R”. Now “R” would become the most used and most ambiguous word in the Loglan vocabulary. Further, if you can make the very subtle semantic distinctions necessary to produce 17 or 22 interpretations of “pretty little girls’ school”, then you should [be able] to make the very obvious distinctions in the potential meanings of “R”. [Thank you for your vote of confidence, soi drani crano, but I’d rather you did it; the exercise would be good for you, soi clafo! You obviously have enough skill and interest in this topic to pursue it productively...that is, to all our advantage.]
To other logli let me say that the Keugru will welcome any criticism of, or alternatives to, the ideas we have recently been generating on this intriguing topic ... most of which have already been bruited to the logli list, I believe.
I suppose that “the Keugru” is the Loglan Academy. [Yes; the word keugru is derived from the metaphor kerju grupa = caretaking-group = academy... which is still a fifth type of metaphor, the “activity + actor” type!] I am glad to hear that it (or they or him) will “welcome any criticism of, or alternatives to, the ideas ...”.
“Bruited to the logli list”? Is “bruited” a nearly archaic English verb or a Loglan word with suffixed “-ed”? I find the meaning of this phrase obscure. [Bruit means to make public/noise abroad, and is an old but not “archaic” word. What sounds or looks “archaic” to a listener/reader is, I suppose, what he or she is not accustomed to hearing or seeing but assumes once had a well-defined meaning. I do confess that I have the logophile’s interest in keeping lovely old words (like logophile and bruit) alive by using them now and then. The designation logli list refers to the e-connected, computer-using group of active logli who write each other about Loglan over the computer networks and among whom such problems are often “bruited” and then discussed.] I trust that your “welcome” implies a two-way dialog. I’m listening [and I trust you’ve heard, soi crano].
This letter, though in places acerbic, was fun to deal with. I’ll make sure that, throughout our interchange, Ex-Member Doudna receives complimentary copies of the issues of Lognet in which his remarks appear. Maybe we can engage his support again...put him in charge of the metaphor-typing problem, soi srisu, in which he evidently still has a lively interest.
Finally, I should comment on Paul Doudna’s perception that we have an “inner circle” and that it’s “overly secretive and exclusivist.” We do have one; every organization which depends on volunteer work splits into active and inactive parts. So P is correct about that: there are a group who regularly volunteer to do L work. At the moment our Tiftuagru (Offering-Worker-Group) contains ten members. But it’s not exclusive. Anyone who wishes to join by sharing in our work, may do so. Our doors are open. As to being secretive, we try hard not to be. We publish the discussions and decisions of our work-groups in the widest possible way. E.g., here. — Hue Djim Braon