Eliminating verbal adjuncts

[circeus]

As a feature I want to be in Mfalen, I want to try building it without verbal adjuncts. That is, nouns are either part of a noun phrase, or one of the maximum three arguments a verb can have (Mfalen verbs are explicitly marked for valency). I've already figured out how this affect valency-increasing operations like causatives¹, as well as locatives² and instrumentals or adverbs of manner (fairly easily converted to nominal adjuncts).

Am I overlooking any obvious constructions I will need to deal with down the road?

(1)Basically, a Mfalen causative (for example) expresses the causer, but cannot express the agent, similar to how many languages cannot express the agent of a passive
(2)With a few exceptions (i.e. meteorological verbs), locatives cannot even be arguments, and technically classify as adverbs, but I like the idea of treating the noun as central to the grammar the same way I do for word derivation.

[Creyeditor]

I have to admit that I do not yet fully understand what you are trying to eliminate. Could you give an example in English (or some other natlang)? Do you want to eliminate some kinds of (verb-like) dependent clauses, like infinitive constructions, raising, control, etc? Or do you want to exclude nouns that are adjuncts to a verb phrase, like locative expressions, time expressions, etc? Just to clarify. Maybe I will be able to answer your question then.

[Ares Land]

As a feature I want to be in Mfalen, I want to try building it without verbal adjuncts. That is, nouns are either part of a noun phrase, or one of the maximum three arguments a verb can have (Mfalen verbs are explicitly marked for valency). I've already figured out how this affect valency-increasing operations like causatives¹, as well as locatives² and instrumentals or adverbs of manner (fairly easily converted to nominal adjuncts).

Am I overlooking any obvious constructions I will need to deal with down the road?

(1)Basically, a Mfalen causative (for example) expresses the causer, but cannot express the agent, similar to how many languages cannot express the agent of a passive
(2)With a few exceptions (i.e. meteorological verbs), locatives cannot even be arguments, and technically classify as adverbs, but I like the idea of treating the noun as central to the grammar the same way I do for word derivation.

I'd love to see a few examples, especially with locatives
How do you say: I'm in my room, or I'm going to Rome. ?

Wikipedia suggests more cases:
cause, concession (i.e.: although, even though, in spite of ), condition, effects, results, purpose, quantity, manner, state, frequency...

I suppose cause, concession, condition could attach to the subject; effects,, results, quantity to the object...

[Vardelm]

I'd love to see a few examples, especially with locatives
How do you say: I'm in my room, or I'm going to Rome. ?

This seems pretty straightforward to me. The locative would just be the object of a verb meaning "to go to" or "to be in". In the case of "go", a valence/transitivity affix could be used.

[bradrn]

Quite a few languages use locative verbs in serial verb constructions for locatives, which could do what you want. So essentially, the equivalent of ‘I’m going to Rome’ would be something like ‘I go be.in Rome’. Occasionally you also see transitive motion verbs, which would make it even simpler: ‘I go Rome’. (Don’t have time to go into details now, sorry, though I can try find a few good examples later.)

[Richard W]

Quite a few languages use locative verbs in serial verb constructions for locatives, which could do what you want. So essentially, the equivalent of ‘I’m going to Rome’ would be something like ‘I go be.in Rome’. Occasionally you also see transitive motion verbs, which would make it even simpler: ‘I go Rome’. (Don’t have time to go into details now, sorry, though I can try find a few good examples later.)

Transitive verbs of motion seem quite common. The IE accusative of direction is very similar to an the accusative of the direct object (as in the Latin example above - eo Romam), and Pali passive participles can be used for the place gone to. The Thai verb for "go" works as though transitive, as does the verb "to part from", often translated as though a preposition meaning "from".

[circeus]

Quite a few languages use locative verbs in serial verb constructions for locatives, which could do what you want. So essentially, the equivalent of ‘I’m going to Rome’ would be something like ‘I go be.in Rome’. Occasionally you also see transitive motion verbs, which would make it even simpler: ‘I go Rome’. (Don’t have time to go into details now, sorry, though I can try find a few good examples later.)

Transitive verbs of motion seem quite common. The IE accusative of direction is very similar to an the accusative of the direct object (as in the Latin example above - eo Romam), and Pali passive participles can be used for the place gone to. The Thai verb for "go" works as though transitive, as does the verb "to part from", often translated as though a preposition meaning "from".

French "Aller" also works that way hence why the infinitive goes "Allons-y" or "va-t-en": the bare verbal forms are entirely ungrammatical (in fact, ALC, p. 156 has an error beccause of of that, the French example should read Si nous y allions?).

I'd love to see a few examples, especially with locatives
How do you say: I'm in my room, or I'm going to Rome. ?

Mfalen will not have a copula, but it will have a locative verb (though it will be used only when a distinction marked on the verb is required). Verbs that have an inherent locative component (i.e. go to [place]) are amongst the exception that can take locative as arguments. I'm also entertaining a class of verbal derivations that basically take the Uralic case system and express them as verbs, and then combine that with verbal incorporation so that what in English is "I ran across the field", in French becomes something like "I crossed the field with running" (where "running" is a noun, not a verb form), and in Mfalen you get something like "I run-fieldcrossed" (or "I throw-bedtoped", "I throw [X] on the bed").

Wikipedia suggests more cases:
cause, concession (i.e.: although, even though, in spite of ), condition, effects, results, purpose, quantity, manner, state, frequency...

I suppose cause, concession, condition could attach to the subject; effects,, results, quantity to the object...

I intend to construct the vast majority of those as clauses of some sort. Many arguments that would be adjuncts of some sort in IE languages can be made to fit in the tri-argument structure of Mfalen, especially for verbs that aren't inherently bitransitive to begin with. The language is not shy of parallel constructions where needed either, and argument stripping is more widely available in Mfalen than in English so it's doesn't feel like it weights things down. Other times, they can be expressed through verbal derivation instead. It's worth noting that Mfalen does have adverbs, but as they are a separate word class from nouns, I'm not treating them as adjuncts for this argument.

[bradrn]

OK, I have a bit more time now, so as I promised, I’ll go into a bit more detail about the stuff I mentioned in my last post here.

Essentially, the main idea is to eliminate adjuncts by turning them into the objects of serial verbs. A typical example is instrumentals, which are very often expressed by verbs such as ‘take’ or ‘use’:

‘1s take knife cut bread’ = I cut the bread with the knife
‘1s cut bread use knife’ = I cut the bread with the knife

Similar constructions can be used to express comitatives and benefactives:

‘1s accompany 2s walk’ = I walk with you
‘1s help 2s make phone.call’ = I’ll make a phone call for you

Causatives often use ‘give’ or ‘make’:

‘1s give 2s go’ = I make you go

Additionally, SVCs often express notions of motion, location and path. Typically the locative element is expressed as a verb within the SVC. For instance:

‘1s take come 3s’ = I bring it here
‘1s take go 3s’ = I take it there
‘3s descend return’ = (S)he comes down
(Sougb, Reesink 2002) yen y-aiga duhu=da ‘you 2p-cross water=go’ = Cross the river
(Abui, Kratochvil 2007) di=ng wahai mara ‘3=see look go.up’ = He looks up
(Maybrat, Dol 2007) t-tu aya m-amo cerek a ‘1s-pour water 3-go thermos Q’ = Should I pour water into the thermos flask?

I think that covers the vast majority of ‘verb adjuncts’, as you’ve been calling them; please tell me if I’ve missed any!

EDIT: I just found a particularly impressive SVC from Cantonese (Matthews 2006), with several adjuncts in one sentence:

keoi⁵ gon² faan¹ lai⁴ wan² jan² heoi³ bong¹ lei⁵ waan⁴ cin² bei² ngan⁴hong⁴
3s rush return come seek person go help you return money give bank
He came rushing back looking for someone to help you pay back the money to the bank

[Ares Land]

Mfalen will not have a copula, but it will have a locative verb (though it will be used only when a distinction marked on the verb is required). Verbs that have an inherent locative component (i.e. go to [place]) are amongst the exception that can take locative as arguments. I'm also entertaining a class of verbal derivations that basically take the Uralic case system and express them as verbs, and then combine that with verbal incorporation so that what in English is "I ran across the field", in French becomes something like "I crossed the field with running" (where "running" is a noun, not a verb form), and in Mfalen you get something like "I run-fieldcrossed" (or "I throw-bedtoped", "I throw [X] on the bed").

OK! That sounds very promising.

d. It's worth noting that Mfalen does have adverbs, but as they are a separate word class from nouns, I'm not treating them as adjuncts for this argument.

Ah, yes, I was wondering if adverbs counted.

[Kuchigakatai]

(1)Basically, a Mfalen causative (for example) expresses the causer, but cannot express the agent, similar to how many languages cannot express the agent of a passive

This seems pretty surprising... So you can't say 'her son' in 'I made her son laugh'? How would you say something like this? Or 'I made her son finish his food'.

If you have ditransitives you could try making the agent an indirect object, and some natlangs do that, just saying...
I made [to her son]_INDIRECT-OBJ [finish his food]_DIRECT-OBJ

[vegfarandi]

How would you express a complex set of multiple locatives without the ability for an arbitrary number of adjuncts?

I'm going to Trastevere in Rome on the road from Florence.

Things like valency increasing operations usually don't allow arbitrary number of new arguments, cross-linguistically there seems to be a cap of roughly 4-5 arguments per verb.

[bradrn]

How would you express a complex set of multiple locatives without the ability for an arbitrary number of adjuncts?

I'm going to Trastevere in Rome on the road from Florence.

Things like valency increasing operations usually don't allow arbitrary number of new arguments, cross-linguistically there seems to be a cap of roughly 4-5 arguments per verb.

Can’t speak for Circeus, but serialising languages often do it something like this:

1s come.from Florence go.along road go.to Trastevere be.in Rome

That is, in literal translation: ‘I come from Florence, go along the road, go to Trastevere, ending up in Rome’.

[circeus]

Hey guys. There's been a lot of very interesting points and I do intend to discuss my ideas further in response to them, especially regarding complex locatives. However, ADHD is (unfortunately <<;;;) a thing that exist and it seems like it grips me every time I intend to answer. Especially since I actually had an answer typed up back Monday, but somehow never hit the send button for it!

I'll throw in that on my favorite aspect of attaching locations and instrumentals to individual verbal arguments is that it makes it so simple to assign separate ones to different participants in an action, something that requires some hoop jumping in all languages I know of enough to make a comparison to.

[bradrn]

I'll throw in that on my favorite aspect of attaching locations and instrumentals to individual verbal arguments is that it makes it so simple to assign separate ones to different participants in an action, something that requires some hoop jumping in all languages I know of enough to make a comparison to.

I don’t really understand what you’re saying here — could you give an example?

[circeus]

I'll throw in that on my favorite aspect of attaching locations and instrumentals to individual verbal arguments is that it makes it so simple to assign separate ones to different participants in an action, something that requires some hoop jumping in all languages I know of enough to make a comparison to.

I don’t really understand what you’re saying here — could you give an example?

basically, in a sentence like "I fought Zompist", specifying different weapons for each participants requires some degree of contortion because typically languages (caveat: those I have some degree of familiarity with) tend to assign the instrument to one favored argument, usually the verbal subject. In Mfalen, you end up saying essentially [ I with a sword] found [Zompist with an axe]. Similarly with location (although a lot of languages would merely express the difference in locations through entirely different processes instead).

I'm aware this works for nouns ("a warrior with an X", "the man on the ground") but it typically breaks down to some degree when you try to say the same thing with pronouns or proper nouns ("I fought from the ground Zompist [*atop a wall]"). English will soemtime let you move the adjunct to the syntactic subject, but you can't readily add another and assign it to the syntactic object: "I, with a sword, fought Zompist" works, but you'll have to introduce some extra syntactic luggage to specify Zompist's weapon.

"zompist and his axe" admittedly works, but this repurposing of coordination seems like a different beast in this case for some reason (plus it's nearly one am, so you'll excuse if I don't elaborate any further for fear of tying myself into a knot with my own ideas XD). I figured it out: the reason it wouldn't work for Mfalen is because it implicitly makes the instrument animate.

[circeus]

Especially since I actually had an answer typed up back Monday, but somehow never hit the send button for it!

Turns out I don't have that much to say because, for some reason, I thought the October 3 post had never been actually sent.

(1)Basically, a Mfalen causative (for example) expresses the causer, but cannot express the agent, similar to how many languages cannot express the agent of a passive

This seems pretty surprising... So you can't say 'her son' in 'I made her son laugh'? How would you say something like this? Or 'I made her son finish his food'.

If you have ditransitives you could try making the agent an indirect object, and some natlangs do that, just saying...
I made [to her son]_INDIRECT-OBJ [finish his food]_DIRECT-OBJ

Much the same way as languages that cannot express the agent of a passive do: repeating the verb. It's not a superfluity either, as I was a big fan of the Nahuatl difrasismos as soon as I read about them years ago, as well as Nahuatl's penchant for stringing multiple verbs meaning pretty much the same thing. So having to say "he finished his food; I made [him] do it" fits right into what I want typical Mfalenstyle to be like. Heck, in fact Mfalen allows "He finished and [∅] made finish his food" i.e. "He was made to finish his food". ([∅] being a subject marked on the verb as suppressed.)

I'm not yet entirely clear how much object clauses will factor in Mfalen outside of reported speech, but since the causative will not be compositional, it is not relevant here. Apologies for not being clear that Mfalen's causative is formed via an affix, not a syntactically like in most IE languages.

How would you express a complex set of multiple locatives without the ability for an arbitrary number of adjuncts?

I'm going to Trastevere in Rome on the road from Florence.

Depends whether these are explicitly independent or forming a single locative, really. In the former case, they would probably represent each a separate verb of the type I described earlier. In the latter case, they would simply form the object of the "go" verb as noted above.

[Kuchigakatai]

Much the same way as languages that cannot express the agent of a passive do: repeating the verb. It's not a superfluity either, as I was a big fan of the Nahuatl difrasismos as soon as I read about them years ago, as well as Nahuatl's penchant for stringing multiple verbs meaning pretty much the same thing. So having to say "he finished his food; I made [him] do it" fits right into what I want typical Mfalenstyle to be like. Heck, in fact Mfalen allows "He finished and [∅] made finish his food" i.e. "He was made to finish his food". ([∅] being a subject marked on the verb as suppressed.)

Ah, I see. That makes good sense.

By the way, Classical Arabic couldn't express the agent of a passive either, but it didn't solve it by repeating the verb. It just used VOS order, while keeping the object (O) accusative and subject (S) nominative.

This is also true of modern Spanish, which doesn't express agent NPs of verbs in passive constructions, except for the most formal layer where the old passive construction of Late Latin (be + past participle) is still in some use (which express the agent with the preposition por). Much like in Classical Arabic, you just use OVS order instead, the agent still remaining the subject S of V, after fronting the O to the beginning.

[Al paciente]_OBJ lo mataron [sus propios doctores.]_SUBJ
'The patient was killed by his own doctors.'
(mataron is in a 3PL form, agreeing with subject sus propios doctores 'his own doctors')



(If there was no expressed agent, then Classical Arabic simply used its passive conjugations to which an agent NP couldn't be added, or the "form VII" mediopassive. And modern Spanish uses its normal passive constructions: the passive reflexive construction, the 3SG impersonal, or the 3PL impersonal.)