-
Notifications
You must be signed in to change notification settings - Fork 4
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
RFC 0003: Regarding Cs #5
Comments
So, Opening Statement: the VxCs proposal was a sort-of afterthought. While I still believe that it’s entirely feasible to get by with the second row of cases + root-derived Cs, the idea—as I presented it—is rough. If we work something sensible out, we can file a new RFC. To recap, my main gripe with VxC(s) affixes is that whenever one wishes to include such an affix in a formative of theirs, they must go through this ‘let’s see if I remember this correctly’ moment, where they check:
This would be allowed: ‘Affix degrees are replaced with Case accessors (either the second row of Cases or all Cases); save for a handful of grammatically prevalent operator-like affixes (such as COO and TPF, which it should be noted may at any time be recast into Categories of their own), all consonantal affix values name the same semantic content as that of the corresponding root values.’ Of course, a design which minimizes the amount of irregularity required to get by pragmatically is most desirable.
You really think that’s the only problem humanity’s got to solve? lol (Jokes aside: there’s a whole gradient of irregularity across the affix lexicon. It’s enough that the ‘linear gradient’ types of affixes—such as -x—are already irregular in a way, since Degrees 1 and 9 represent excess.)
I think it’s clever in that it doesn’t require any more of the language to provision—i.e., no need for a separate part-of-speech that is the conjunction. (Though both in 2011 and in 2020, you can extract the COO into an adjunct, and then you can pretend you’re a natlang.)
It works for an overwhelming majority of cases, just like 2011’s Schematic Format covers 90% of Format usage. Agreed, though, 90% ≠ 100%, so on to your suggestion(s):
Watch your mouth ;) Lojban’s {fi’o} is utter shit, and its sole existence shows the profound weakness that lurks within the BAI. But moving on: [My myself’s note: I’m responding sequentially, so, since right now I’m yet to read your proposal in full, but have seen trimmings, I can say that I’m excited. Now imagine layering a temporal axis on this post that goes along it lengthwise, from the top to the bottom, hehe.]
I’d say:
Quite the opposite: intransitive verbs in Ithkuil can be ABS or IND.
So if this is the reason why you decided to use Latejami’s case roles rather than Ithkuil’s, and my reasoning above is correct, then I do recommend returning to the appropriate Ithkuil cases: THM, ABS, ERG, IND. Regardless, I liked the parabole. :)
Without Designation?
This is interesting… I totally overlooked the fact that in many cases, we’re dealing with something resembling a transistor (physically). I see where your proposal is going, so let me propose a clearer version: For base meaning Your example, if copied over to my interpretation, would be Thoughts? |
Argh, I thought we said to JQ not to do this
Yes, and I think it's cool of an ihtkuilic language to have overprecise suffixes, like
This argument is good by itself, but
(many are still horrible, like -kc Edible or Autonomous Plant Parts/Components, and I just randomly scrolled and picked the first I've seen)
Yeah I really think that. You can read it in my post, I've precisely used the words I said that, because TPF requires special parsing rules, and no semantic interpretation will ever change that. As I've said, this is not the case for COO.
Agreed x)
Worse, some suffixes break the gradient, like -t, which should be type C
My point is that we don't need that much semantic precision. My proposal could be rephrased that way: we need morpho-phonological space, we don't need 81 cases worth of precision, and Latejami proposes a model were only 4 cases are needed. Let's use that, and use the remaining space for showing Type, etc.
I’d say: LG-(verb) this-ABS two-THM second-PAR, or «algalá rre aksal urvalui».
Yeah I agree
So ABS always imply an agent. So I don't disagree with
but this misses the point This is problematic for me, which is why I used AFF, even if AFF has also an incorrect semantic..
So no. THM, ABS, ERG, IND 's semantic are too precise FOR MERGING Cs and Cr (for a normal use as a clause-case, they provide interesting precisions that give to Ithkuil a part of its charm). For the goal of merging Cs and Cr, they should be generalized for ABS and ERG, and specified for THM:
In the STEP THREE: Analysis/Derivation Using Ithkuil Morphological Categories of the translation of the opening Line from Anna Karenina, you can read that
But here, following Latejami's semantic, "family" should be Patient, because one of its caracteristic that the families endure are described. In Latejami, THM is always related to the Patient (at least, also the Agent for action verbs)
ah yes
Correct. Your explanation is indeed far more clearer than mine, but it does not cover everything The "returned / applied" distinction covers the fact, that, either the entry case or the exit case (more precisely the result of applying B and R with cases C and D) can be the new B for the outer scope.
Final words: my proposal is just the same as your, but with latejami's roles for efficiency. But more generally, the definition of ERG and ABS should be changed, it is bad that ABS always imply an Agent, as it makes it too precise. |
I like the extensive case list, though I think it could be redesigned with case-accessors & their inverses in mind. The difference between EFF and ERG is important. (For instance, it marks the difference between manslaughter and murder.) And if case is doing even more heavy lifting (by being one of the ways we instantiate stems from roots), all the more reason to make it more powerful. Perhaps we don't need 70 of them, but there are distinctions which are meaningful in human life and culture that I think are worth pointing out with case. |
@melopee:
The over-precision can be expressed in the lexicon. Like the emotion roots. Like ‘to do something unpleasant because of social obligation’. (Also note that that’s ‘social obligation’ + some personal perspective affix.)
A rational amount of them—okay. But some offenders are, like wombat likes to point out, extremely Anglo-centric, only serving as a nifty shortcut in English translations. (Then we have people coming to us and saying shit like ‘wOW, ithKUIL is SO inFORmAtION deNSe!!’. Exactly what benchmark-driven development means.) And extending this point, in general, to have affixes whose degrees you can’t express natively without using those affixes—that’s the definition of hackish. JQ uses them as a quick-’n’-dirty hole patching tool. As if roots weren’t enough!…
That’s what I meant. Sorry it didn’t convey itself too well through the joke. And there are far more irregular affixes—it’s just that some are more dissatisfying than others.
Yes. But there should be abridgement. And self-restraint. JQ doesn’t possess these traits, I’m afraid. [Full-time shitting and pissing on JQ—how are we doing?]
Ithkuil greatly benefits from this sort of precision because prepositional phrases don’t translate too well to Ithkuil (they require a whole separate formative, which it is then somewhat hard to connect to with the complement of the prepositional phrase… It’s a mess.) On that note, we might try adverbial affixes: those which determine the semantic relationship between the formative and the clause it’s enclosed within—displacing Case. Now this would rock!
Several problems with this:
But indeed, it’s an irrelevant aside.
I dare disagree with that assessment. Moreover, they might be—to a degree—sufficient. And a carefully cherry-picked handful of Cases might be able to cover all feasible use cases.
Also, I should get this in real quick: CTE is what you may be looking for.
Yes, for a ‘degree of size’ kind of predicate, you would use the CONTENT semantic role—Thematic Case—and then Some Other Case for the measuree. And the Absolutive Case fits here nigh perfectly.
This is fair (outside of VxCs discussion).
This is not okay.
In addition, DAT will be absolutely useless for affixes. It’s already rarely used in conversation. (AGOI suggests DAT for possessives, but that’s just wrong—it’s as if the existence of a book was meant for me, and circumstantial possession is less than that. In 2020, regular Possessive Cases should be fine—as in,
If I understand correctly: this can be handled with CTE.
Yeah…
How about this:
Also, this should become an RFC of its own, nay?
Yes! Totally. |
It is important, but do we really need it for just combining two roots into a suffix?
I hope we will not run short of root space then.. and I don't consider using 4-consonants cluster as a solution
Yeah, I particularly hate the linguistic hedge things
Like fiho or toaq's preposition
I haven't reread the website in a while, but in https://www.reddit.com/r/Ithkuil/comments/9w08ef/lexicosemantics_of_verbs_in_the_new_language_im/ you can find
No, at least for what I understand. For me, CTE is a hook for the verb as a whole. It is the "verb case" you talked to me about.
would mean
and not "you are blue" (which is what I call an existential state)
Yes, but this is not what I propose. I propose to
Yes, but you can coherently generalize ABS into containing the meaning of DAT, like Latejami does (relevant section: http://rickmor.x10.mx/lexical_semantics.html#S4_3_8)
Agreed, which is one more reason to merge it with ABS into a new Patient case with generalize semantic
Yeah that's just a bloat
How is that? After rethinking about this, there might be a real use for this feature: let's try with
would mean "two seconds" but what would mean
and
? I thought this was a clean way to distinguish Modular from Casual suffix application (skewers's terminology for the other who read that)
Or V(y/w)V, but we will decide the general phonoaesthetic of the morphopho later (NB: I loathe glottal stops, having one per word between two V is already enough for me)
Okay so seems we agree on the general idea.
e.g.
I had that hunch before, but now it very clear that the result will be extremely verbose. So, conclusion of my conclusion:
Yeah |
Agreed. There is some balancing finesse going on here—and I wouldn’t mind merging case roles in order to hit that sweet spot. However, I consider the best solution to be one where a small set of cases gets short-and-easy forms, and all other combinations of cases—possibly not all cases, but a wider set either way—get composite forms. At the moment, I’m considering the set {CTE, THM, ABS, ERG, IND}, Cartesian-producted with itself, yielding 25 forms (that’s little in the world of TNIL—we have 36 ‘simple’ V values already, and they may exist more), and then that set × all cases, and all cases × that set, using some V’V shenanigans.
Ditto.
The Affective Case is messed up. 😠 Let’s have it redefined clearly for Freetnil, eh?
I disagree. With the right semantics for AFF—involuntary, affective experience—we can let go of it for affixes altogether.
I’ve considered that already. 😎 The second Case (exit Case) is the one that’s returned, always, so
Sort of: when the entry Case isn’t the same as the exit Case, then we’re taking out something else than what we’re putting in—Modular. And when the entry and exit Case is one and the same, we sort of pass the base meaning through an additional assertion—Casual. [Anybody not familiar with Skewers: disregard this.]
That’s more or less what my secret plan was. I plan to begin this kind of work soon (it’d be a pain to do it as a group, so please leave it to me for now). (Your decomposition is good—well, almost, because there should be some atomic way to convert an abstract unit—like ‘year’—to that which, when measured in its terms, is the unit—‘something which lasts for a year’. This would be wonderful as a Case—I may add that in soon; it’d also offset the semantic gap created through removing the Objective specification from the duration roots—which meant exactly the above.) ((Edit: Then again, the example that you gave is sort of cheaty, as is Ithkuil’s handling of orders of magnitude. If we don’t make numbers in general succinct, the corresponding suffixes won’t be either.))
Eh. An RFC is in the works. I hope that I present my case well. Until then! |
(Oh, I missed this:
Yes. In logicspeak, the examples above would be (with CTE is essential to my affix concept since it gives you a means of omitting a Case access. Which is useful at times: This really needs clarifying in the RFC.) |
I would use THM here, not CTE |
@melopee:
That actually depends on the content (pun intended). If by ‘feeling’ you mean ‘state of feeling’, you should use CTE; if you mean ‘the internal qualia/content that is felt’, you should indeed use THM. I can only blame myself for giving a bad example. Here’s a better one:
I’d agree. But there's a cool distinction to be made here: ‘I love you for your body’ would have ‘your body’ |
Also, yesterday, @porpoiseless and I discussed a potential ‘complement’ Case which would tag a secondary theme, complementing the primary theme (THM) without physical involvement (X is related to Y, X is a measure of Y, etc.). This may cure our ABS vs. AFF troubles. |
I don't understand how. Plus this is pretty similar to the new meaning of STM
No, this example is problematic, since now THM has two meanings that should be split, IMO:
So "my love for you"
I would rename your no-op case CTE to sth else, and use CTE for my "THM content", and keep "THM" for my THM focus.
I don't understand anything here. How is "I'd agree" related to what I've said?
my goal was to highlight that the two uses of THM
might be incoherent. For me, "sth" in "I feel sth" should be what I called "THM content" Moreover, why do you use AFF for you in ‘I love you for your body’ |
‘2 km is the size of this mountain’ would have ‘2 km’ in the THM and ‘this mountain’ in this new case. It’s a bad idea to use physical cases for describing incorporeal interrelations. JQ’s new STM is just a dirty patch.
I don’t understand the context. ‘Love’ would be the verb/head of the noun phrase, would it not? The focus is either the content or some other case. It’s not a case in and of itself. Besides, take a look at what Morphology says:
That makes it pretty clear, I think, and also why we should have a ‘second’ CONTENT.
In the first case, THM is the content of the emotion. In the second case, THM is the emotion itself. Nothing wrong in here—this is how those predicates are defined.
My bad: I meant STM. |
Ah, ok. I hope then that the distinction between the primary and the secondary THM will be clear for each root then (e.g. here that secondary THM tags the amount of size for the root "is.size.of")
Yeah, but with a "THM-content" case-accessor. It will be clearer with a full proposition, even if it's a bit weird in english: "I love you with intensity" —›
THM-content is intended to refer to content of the state/action. Not the action seen as an event or as a spatial/temporal reification/activization, but more as what it consists of from an abstract point of view. This is more precise than your CTE, which tags the meaning of the root.
:) With my personal conlang experiment, I noticed that I didin't have the same use for Cases than Ithkuil. |
It is already somewhat clear from the semantics of each root—‘be a degree of size’ has the degree of size as its primary semantic content.
This makes no sense to me. How is what you’re describing different from the new Essential case? And then, your example suggests the use of Functive…
This does make sense. Take a look:
(To have ‘love’ in ‘to feel love’ be the CONTENT would make the that root redundant to the ‘to feel an emotion’ root. We assume that specialized roots take on the semantics of their specializations, at least in a majority of cases.) If you wanted to use the former root—‘to experience an emotion’—to say something like ‘I feel love for you’, you could do something like
But this is pointless, as ‘love’ is an emotion/affective experience already [1], so you could just say
and skip the awkward ‘THM* of THM’ bit. (Any instance of [1] Because of this, you could even say
—literally, ‘my affective experience is the love for you’. Or, to highlight the redundancy in the example: ‘my affective experience is an affective experience of love towards you’. Drink responsibly. |
Yeah, but I don't really like your Essential case. My THM-content is like Essential but without the "reason / profound cause behind sth" part. It is also different from the Functive in that it does not tag the whole event, it just specifies the content of the event, what the event is "made of" from an abstract point of view.
Yeah this makes sense, but this is not how I envision case. Even if I do think your approach is more pragmatic and will be more useful in ithkuil |
Eh, I think my Essential is as little ill-defined that you can get. ‘Reason / profound cause’ is only a partial description; the RFC reads ‘ideal/intended cause-and-purpose or profound meaning’. Essential concerns the fundamental state/action that the base state is an example, manifestation, realization, of actualized means of. So for ‘feel love’, ‘have affection for/need to care for’; for ‘eat’—‘nourish’; for ‘clothing as self-expression’—‘self-expression’; for ‘set on fire’—‘make heat’. For some X’s, SNT-X is contained within its cause, or within its purpose; for others, it’s more complicated. Now go ahead and define your ‘THM-content’ (please find a better name for it).
How do you envision Case? |
More like accessors to a common argument structure shared by all verbs. Not that this argument structure is inherent or natural, it should be designed when creating a conlang.
I don't feel a need for it, as I initially envisioned to use CTE; but if you insist, let call it "GiST".
This is not a case-frame.
Here, keep in mind that "-THM" in "love-FRAMED-THM" marks the role of the case frame in the main clause. It does not access any argument of "love". In fact, this GST in kinda like SNT, maybe they are the same. But I wasn't sure since "Reason / profound cause" seems important in SNT's definition. On the contrary, I thought of GST as opposed to Methodic and Material. |
Now I don’t understand how this
I feel like I should bring this up again: the definition is ‘ideal/intended cause-and-purpose or profound meaning’. There is no focus on the reason part—not any more than there is on the result part. Grr.
I really wish this is how this worked, but it’s impossible. It relates to an idea that I’ve been flaunting recently: that some case roles in a relation are essential to it. What this amounts to is that:
In other words, root concepts bind case roles because their wide semantics fits the narrow semantics of the concepts. Sadly, this is lexical information; without it, things like |
Is it that bad?
Consider the reverse: all affixes are perfectly regular. This makes Vx pretty useless as a slot. One important vocalic slot is wasted just for precising the degree the affix.
Even if it's more irregular, I think it's more efficient (from a cognitive standpoint) to:
Is it that bad?
Some VxCs are pretty interesting, like PWF (part to whole functional metaphor).
It seems convenient to me to be able to quickly derivate such part to whole metaphors.
Though I agree with you, many VxCs are just bag-of-words at best, and do not seem useful at all.
Okay this is bad.
I agree, but the only syntactically irregular affix is, IMHO, TPF. (DCD and SQP are weird, but do not introduce new syntax, it changes only the meaning of the surrounding syntax, so this is not as bad as TPF)
As for COO and friends, I don't worship with JQ's adverbial interpretation of connectives, but it is coherent and interesting.
Agreed. We need a general mean to make Cr and Cs work together.
Even if, as I've said above, I don't think removing all suffixes with irregular gradients will make the language better / more usable..
Anyway I'm also dubious about the solution you propose replacing Degree by Case-Acc'.
I don't think it will work because in order to be morpho-phonologically efficient, we have to use very few structural morphemes. But using Case takes a lot of space..
Moreover, Case has too much features for what we need.
We don't need 70 Cases for just applying functions to other functions, and gaining a new function (an affix) as the result.
Using all the power of Case can provide handy shortcuts, but morpho-phonologically, I don't think it will pass.
Instead, for combining Cr and Cs, I propose to make a new, trimmed version of Case based on Latejami.
Latejami only considers 4 roles: Agent, Patient, Focus (THM), and Agent-Patient (IND), but, contrary to Ithkuil, expands a bit their semantic, and rely on a powerful case-tag system for creating other roles (like Lojban's {fiho})
Let's work through an example, and build the suffix:
with
I assume that -LG- can be used that way:
ABS is not appropriate since it implies an agent.
On the contrary, Latejami's Patient is more general and contains both ABS and AFF.
Let's first define a new Vx Slot, that will show, among other things:
So
-rc DEG4
is modelled as:where
P-return
is a value of this new Vx which shows that the thing to which this suffix is appliedwhere
F-applied-operator
is a value of this new Vx which shows thatRV-(Stem:1)
is the Focus argument ofLG-(Stem:1)
, the "operator" (verb) of this suffixWith only four cases to consider, there will be enough morphological space to show other interesting things like Type or Stem (and maybe other things like skewers's Closure).
However, reforming VxCs as we currently know them will be a massive task
Many of them are based on simple concepts (they don't require the trick described above based on Latejami's), e.g. FRQ; so Cr roots should have clear semantics right from the start if we want to use them as Cs.
For those that could use the Latejami's trick described above, things will get messy quickly. For example, it is possible to generate the suffix RPN Deg1 "slow-paced repetition at regular intervals" as the application of "... repetition at ... intervals" on "slow" and "regular" but this will not be verbose only if "regular" and "slow" have their dedicated roots and Stems.
The text was updated successfully, but these errors were encountered: