Bill Harris wrote: > I'm glad I wasn't the only one somewhat confused,
You certainly are not. Try searching the Forum archives for uses of the term "verb phrase" (atomically, e.g. [1]). > I agree that Ken and Roger (and perhaps most experienced J'ugglers) are meticulous in their wording This is one of the most gratifying extracurricular benefits of J: participating in these forums has sharpened my English [2]. > I like where you're headed, although I might start ["verbs which are the product"] with "verb phrases which ..."; I see what you're saying, but I was trying to define the concept (left arguments to modifiers) without using the term "verb phrase". Or, I was trying to identify the types of verb phrases that are (immediate) left arguments to modifiers. Raul wrote: > It appears that a verb phrase is something typically built from adverbs > and/or conjunctions, and that hooks and forks are not verb phrases unless > they are contained in parenthesis. > I am not sure how you would reach this conclusion, though, if you did not > already understand how parsing works. As I noted in an earlier message [3], we could make any number of retrospective justifications (knowing how the interpreter acts and building a justification for this out of the DoJ), but this amounts to apologia. What we want is deductive (as opposed to inductive)/prescriptive/predictive rules, such that if we did not already have an implementation of J, we could build one. Take the Dictionary's position regarding name stacking ([4], referenced earlier) as an example. We saw that the retrospective attempt to reconcile the Dictionary (J) with the implementation (j), by suggesting that "proverb" referred to "a name for a verb, adverb, or conjunction", as opposed to the plainer meaning of "a name for a verb", could not stand. Ultimately, this is because we discovered that the Dictionary explicitly defines proverb in the plainer sense, but motivation to locate that text was that the only argument in support of the fancier definition is that is how j implements J, and such an argument begs the question. Similarly to "proverb", we should expect "verb phrase" to have its plain meaning, built in analogy to the English language (which is the design theme of the Dictionary [note the name]). To wit: a phrase that produces a verb. If it didn't, we'd have to worry about e.g. "noun phrase", and then arbitrarily create words that had the plain meanings we desire, just so we could talk about J. > The dictionary uses "verb phrase" on two pages: > http://www.jsoftware.com/help/dictionary/dicte.htm > http://www.jsoftware.com/help/dictionary/intro08.htm I personally would try to avoid using this latter page as a basis for formal arguments about J. While certainly it could lend credibility or strength to a J argument, given that it was written by Ken specifically to introduce J, it is not reference material, and so could favor illumination over precision (i.e omit relevant detail). In general, I try to restrict my formal reasoning about J to the Dictionary proper, which is indexed by the last column of [5], from the top down to the heading "Control Structures". The reason for my restriction to the last column is that only that column is reference material (i.e. is canonical), so nothing else can serve as a basis for formal arguments (informal or pragmatic arguments are a different matter). The reason I stop at "Control Structures" is the sections following that do not assist reasoning about J proper (i.e. qua notation); they either help one reason about a particular implementation of J (e.g. Appendices A, B, C) or do not help formal reasoning about J at all (e.g. References, Acknowledgements). This principle is slightly contradicted by Appendices D and E, which I would argue belong in the Dictionary proper. But Roger is responsible for both, and he does not consider the spelling of an error particularly significant (see e.g. [6]), and the "parts of speech" listing is a recent addition which could be considered a user convenience, as the typeface of the Vocab entries already indicated (indirectly) their part of speech. The thread in [4] again exemplifies this principle, where we found strong support for the plainer definition of "proverb" in some material outside the DoJ, but were not satisfied until we found support in the DoJ proper. -Dan [1] Other people that have found "entire verb phrase preceding" confusing and contradictory: " In my confusion I interpreted this statment to mean that the left argument of the atop conjuction in (#...@#) was #i. not just i. I still do not know what constitutes an "entire verb phrase" for a conjuction." http://www.jsoftware.com/pipermail/general/2002-July/009961.html " But it must be that the verb sequence "f g h" is not a verb phrase (the term seems not to be defined in Dic), because the interpretation [of f g h a] is f g (h a)" http://www.jsoftware.com/pipermail/general/2003-April/014448.html etc: http://www.jsoftware.com/cgi-bin/forumsearch.cgi?all=&exa=verb+phrase&one=&e xc=&add=&sub=&fid=&tim=0&rng=0&dbgn=1&mbgn=1&ybgn=1998&dend=31&mend=12¥d= 2010 [2] By the way, am I a "Juggler"? Are we all "Jugglers"? I was under the impression that JUgglers practiced J on Unix, and e.g. participated in Martin Neitzel's mailing lists (http://juggle.gaertner.de/), and that because I practice J on Windows, I could not claim to be a Juggler (but was a more general "Jer", or something). [3] My earlier message: http://www.jsoftware.com/pipermail/programming/2010-January/017683.html [4] Regarding "proverbs": http://www.jsoftware.com/pipermail/programming/2009-August/015810.html [5] J contents: http://www.jsoftware.com/help/dictionary/contents.htm ; J's canonical reference material is indexed by the last column, up to and including the heading "Control Structures". [6] e.g. "the exact spelling of an error message is not particularly significant" : http://www.jsoftware.com/jwiki/System/Interpreter/Bugs06#domainerrorwhenvalu eerrorexpected ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
