John E Clifford, On 29/09/2012 23:15:
> I have now seen some such exchange, but "explains what's going on"
> seems a very generous overstatement and does not in any way alleviate
> my qualms.
I'm wondering if your iPad conceals quoted material in emails, so that you sometimes miss the context? My Android phone sometimes does that. Do remember that the rest of us usually go to greater or lesser pains to quote the bit we're responding to (and cut out any bits we're not responding to).
With regard to the exchange you were responding to, I'd observed that before I can really understand BNF rules such as Jorge and Mike have produced, I have to first of all sit down and work out a set of generatively equivalent rules of the type that human languages use. At Mike's invitation, I wrote out a set of human-language-type rules that are generatively equivalent(ish) to Jorge's core grammar BNF rules. I agree that Xorban grammar must include rules for binding relations (and not just complement relations), but that is not relevant to what Mike and me were discussing.
You go on to say "It clearly results in sentences which I would consider ungrammatical as well as nonsense". Do give examples of such sentences so that we can consider them. I'm not aware of there being any sentences that are generated but ungrammatical -- with the proviso that in Core Xorban all variables must be bound.
--And.
>
> *From:* And Rosta <and.rosta@hidden.email> *To:*
> engelang@yahoogroups.com *Sent:* Saturday, September 29, 2012 4:47
> PM *Subject:* Re: [engelang] reformulating the core grammar
>
> John, you seem to be responding to my message, but did you read the
> exchange between me and Mike, which explains what's going on?
>
> John E Clifford, On 29/09/2012 22:36:
>>
>>
>> Very impressive. Even shorter than mine for toki pona. And even
>> less informative. I know that you want your grammar totally free
>> from questions of interpretation, but not noting that quantifiers
>> bind variables (while sentential connectives do not for all they
>> look the same) seems to be pushing it a bit. It clearly results in
>> sentences which I would consider ungrammatical as well as
>> nonsense.
>>
>> *From:* And Rosta <and.rosta@hidden.email
>> <mailto:and.rosta%40gmail.com>> *To:* engelang@yahoogroups.com
>> <mailto:engelang%40yahoogroups.com> *Sent:* Saturday, September 29,
>> 2012 1:48 PM *Subject:* [engelang] reformulating the core grammar
>> (was: Re: Xorban: Termsets
>>
>> Mike S., On 24/09/2012 20:40:
>>> On Mon, Sep 24, 2012 at 11:22 AM, And Rosta <and.rosta@hidden.email
>>> <mailto:and.rosta%40gmail.com> <mailto:and.rosta%40gmail.com>
>>> <mailto:and.rosta@hidden.email <mailto:and.rosta%40gmail.com>
>>> <mailto:and.rosta%40gmail.com>>> wrote:
>>>
>>> I understand BNF notation; I just can't translate it at a glance
>>> into something more linguistic (like, say, X' syntax). I can
>>> translate it only laboriously. And because my thinking is based
>>> in linguistics rather than computer science, I can only
>>> understand the syntax once it is translated into linguistic
>>> terms. Human language syntax is based on combinatorial properties
>>> of words, and exocentric phrases, if they exist at all, do not
>>> occur willynilly.
>>>
>>> In practice, this just means that to comprehend your proposals,
>>> I'd have to sit down and translate them into a formalism that
>>> makes sense to me, and that takes more time than I have at the
>>> moment, what with me no longer being on holiday. It doesn't mean
>>> you've been too brief or cryptic. If you did want to translate
>>> into a formalism easy for me to grock, then I'd say: assume, as
>>> far as possible, that (i) there is no syntagmatic relation other
>>> than an asymmetric complement-of relation, and (ii) all nodes are
>>> terminal (or, all phrases are endocentric).
>>>
>>> Unfortunately, while I do want to translate into a formalism
>>> that everyone can grock, due to my own background, I am not sure
>>> I would be able to do that based on the description that you have
>>> provided (and forgive me if I seem obtuse). It may help if you
>>> could take the current official grammar or a part of it and as an
>>> illustration render it in the formalism that you prefer. From
>>> there, I should be able to ask questions and figure out what you
>>> mean, and from there on, I will make a point of rewriting my
>>> ideas for productions and transformations in both forms (if I am
>>> able to). Often it is insightful to write the same thing more
>>> than one way, so I really don't mind doing it twice.
>>
>> OK. Working from these:
>>
>> sentence := illocutionary-operator? formula | interjection
>>
>> formula := simple-formula | modifier formula
>>
>> modifier := unary-operator | binary-operator formula
>>
>> simple-formula := (C C C* | q ( C? V? )* q) VkV
>>
>> unary-operator := ( b | f | d | v | m | n ) VkV
>>
>> binary-operator := ( l | r | s | x | g | j ) VkV
>>
>> illocutionary-operator := c VkV
>>
>> interjection := w VkV
>>
>> VkV := V (k V)* (illocutionary-operator formula | interjection)?
>>
>> I would have these as the following. Not just a different notation.
>> I'm doing this without phrases, but a translation into
>> headed-phrases would be simple.
>>
>> 1. word-classes: Interjection, Unary-op, Binary-op, Formula (=
>> Simple Formula), Formula-root 2. A formula-root is a formula or a
>> word whose complement is a formula-root 3. Interjections and
>> Formula-roots have no complements. 4. Unary-ops have one
>> complement. 5. Binary-ops have two complements. 6. Complements are
>> formula-roots.
>>
>> I include Illoc-op in Unary-op, because it is undesirable that
>> Illoc-ops are restricted to sentence-initial position. However, if
>> it were as in Jorge's rules, then Illoc-ops would have one
>> complement and rule 2 would add "and is not an illoc-op".
>>
>> --And.
>>
>>
>>
>>
>>
>>
>
>
>
>
>
>