bradrn’s scratchpad

[bradrn]

A short guide to motional SVCs

I’ve been trying to draft and re-draft this post for… it must be something like six months by now. What’s made it so difficult, I don’t know. But I’ve gotten tired of it, so I decided to rewrite it in as compact a style as I could manage.

The motional SVC

The basic structure of an SVC denoting motion is: direction – manner – path. (The latter two elements can have variable order.) As mentioned in the previous post, the ‘manner’ is a full verb, though often from a morphologically distinct class. The ‘direction’ and ‘path’ here are taken from the following closed classes of verboids, respectively:

Directional verboid	Definition	Corresponding full verb
wal~wel-	‘go’: motion away from deictic centre	√wal- ‘going’
is~yos(y)-	‘come’: motion towards deictic centre	√is- ‘related to coming’
mum~um-	‘around’: undirected motion, or spinning or turning	none

Path verboid	Definition	Corresponding full verb
-fa~fay	‘up’: upwards path	√f-y- ‘rising’
-tan	‘down’: downwards path	√t- ‘falling’
-fe~feŋ	‘exit’: path out of an object or enclosed space	none
-gi~agi	‘enter’: path into of an object or enclosed space	none
-ndo	‘over’: path crossing the top of an area	none clearly (but √ndo- ‘going around’)
-kom	‘under’: path crossing underneath or through an area	none
-ye	‘return’: reverse of a previous path	√yi- ‘related to return’

For example:

Bsiwelesoʼxtan.
[bzi.we.le.soʔxˈtan]
b-si-wel-es·oʼx-tan
1s-PFV-go.B-run-down.B
I ran down.

[…] ygar naʼsŋan nib fasmumesmasndo, trem iŋsarlsesŋun.
[… jgaɾ.naʔsˈŋan.nib fas.mu.mes.mas.ndo tʰrem.iŋ.saɾ.lzesˈŋun]
[…] ygar naʼsŋan nib fas-mum-es·mas-ndo, trem iŋ-sar-ls·es·ŋun
[…] bird DIST.ADN CONTR.FOC HAB-around.A-fly-over.A, lizard 3p-PROG-search
[… whereas] these birds fly around overhead, looking for lizards.

(In the second sentence a speaker could also use a single SVC: fas-mum-lsesŋun-ndo ‘always looks around overhead for’. Or even fas-mum-esmas-lsesŋun-ndo ‘always flies looking around overhead for’, with two main verbs esmas-lsesŋun ‘fly search’ in the manner slot, though that feels a bit bulky compared to the biclausal construction.)

Not all of these components are necessarily required. Most straightforwardly, the path can be omitted, which increases vagueness but doesn’t cause any more significant change: thus the verbs above would become wal-esoʼx ‘run’, mum-esmas ‘fly around’.

Omitting the manner verb is a little trickier, since at least one full verb is required. For such a case the directional can be replaced by the corresponding full verb: e.g. walen-tan ‘go up’. (If if the manner is already specified as a full verb, it would be very odd to use a full directional verb too: ??walen-esoʼx.) Other motion verbs are also possible, see below. Dialectally, some speakers use an exceptional construction with two verboids and no full verb at all, though this is proscribed: ?mum-ndo ‘go around over’.

The directional is the only element of a motional SVC which cannot be omitted: it is the element which defines that motion is occurring. If it is absent, then the construction becomes an SVC of pure path with no actual motion occurring. For instance:

Bsilsenŋunfayi.
[bzil.sen.ŋunˈfa.ji]
b-si-ls·en·ŋun-fay-i
1s-PFV-look-up.B-3s
I looked up.

Here -fay does not refer to an object moving up, but rather the path of the gaze being directed upwards. There is no actual motion here so no directional is specified.

The resultative motion SVC

A different kind of motion construction is seen in the following:

Mbirwemfeysi?
[mbiɾ.wemˈfej.si]
mbi-rw·em-fe-is-i
2s.Q-take-exit.A-come.A-3s
Did you take it out?

This is an example of a resultative motion construction: the motion occurs after the action and as a direct result of it, rather than simultaneously with it. If the main verb is transitive, the moving item is its object rather than its subject.

The above example has structure action – path – direction. As for motional SVCs the path may be omitted with a slightly different meaning: rwem-is ‘take towards’. It has some flexibility of position, and may be placed before or after the directional verboid with no change of meaning: rwem-is-fe ‘take out’. The directional cannot be omitted when there is motion, as in this case.

The second component may be expanded out into a full motional SVC, though this is less common. For instance:

Siʼisnimparwelenoʼxndo-i.
[siʔ.is.nim.paɾ.we.le.noʔxˈndo.i]
∅-si-isni·m·par-wel-en·oʼx-ndo-i
3s-PFV-throw-go.B-soar-over-3s
S/he threw it soaring overhead.

This Eŋes sentence is roughly as odd as its English translation — acceptable in some circumstances, but oddly over-detailed most of the time. More usually one would use a simpler resultative: siʼisnimparweli ‘s/he threw it out/away’. The motional component is almost always included, though: ?siʼisinimpari is odd in the opposite direction.

Finally, it’s worth mentioning that there are non-motional resultatives as well. I already mentioned one above: esmas-lsesnun ‘fly-search’. These motional resultatives are really just a special case of those, albeit a particularly common one.

---

I’m going to stop here for tonight (so that I actually get something published at least). I still have to cover locative objects, either in their own post or combined with other sorts of non-motional object-adding constructions.

[Lērisama]

I'd like to say that I like this, but don't have the time/energy to give it a proper reading just now.

[bradrn]

I'd like to say that I like this, but don't have the time/energy to give it a proper reading just now.

Thank you!

I completely sympathise with having neither time nor energy. Here’s a summary of the main patterns, for what it’s worth:

• direction – manner verb – (path): motional SVC
• direction verb – (path): motional SVC without manner
• action verb – (path) – direction: resultative motional SVC
• action verb – direction – (path): resultative motional SVC
• action verb – [direction – manner verb – (path)]: nested resultative motional SVC

[Lērisama]

Thank you for the summary. I've had a better read now and I have thoughts:

• I really like how the Eŋes verb complex is shaping up. It reminds me a bit of how I'd like Lēri Ziwi's verbs to be – just about that level of complexity, although marking different things differently.
• Is -yos(y)- suppletive, or related to -is- somehow? It doesn't seem to match the other A/B versions in some way, although the patterns are rather disparate.
• What kind of situations would the nested pattern be used in? From the translations and comparison with the non-nested examples, it seems like when both the action and the resultant motion are important to the situation, but for the ‘Si’isnimparweleno’xndo-i.’ example, I can pretty much only think of some sports¹.
• I feel sorry for the lizards.

¹ I have no idea which, because my sports knowledge can be summarised as atrocious, but there must be some where throwing a ball up high is important, surely

Edit: fixed bolding

[bradrn]

I really like how the Eŋes verb complex is shaping up. It reminds me a bit of how I'd like Lēri Ziwi's verbs to be – just about that level of complexity, although marking different things differently.

On this point, I can recommend Matisoff’s Typology of Polysynthesis. There’s a lot of different ways of making complex verbs, and this is just one of them.

Is -yos(y)- suppletive, or related to -is- somehow? It doesn't seem to match the other A/B versions in some way, although the patterns are rather disparate.

Yes, the protoforms are *yusaa- vs *yusaye-. Syncope works in weird ways. (See also: Old Irish.)

What kind of situations would the nested pattern be used in? From the translations and comparison with the non-nested examples, it seems like when both the action and the resultant motion are important to the situation, but for the ‘Si’isnimparweleno’xndo-i.’ example, I can pretty much only think of some sports¹.

…maybe if you wanted an especially vivid description of throwing a boomerang? I don’t really know, and it’s a rather contrived example.

More broadly, nested SVCs are reasonably common in general, both in Eŋes and in some natlangs. Even if this particular construction is rarely used, it’s a good introduction to the type.

I feel sorry for the lizards.

So do I, but the birds have got to eat something…

(It’s not like this is made up, even. My perfectionism meant that I actually researched birds of prey to work out what a sensible object for that sentence would be. Apparently lizards are a major component of the diet of the wedge-tailed eagle, for instance.)

[Lērisama]

I really like how the Eŋes verb complex is shaping up. It reminds me a bit of how I'd like Lēri Ziwi's verbs to be – just about that level of complexity, although marking different things differently.

On this point, I can recommend Matisoff’s Typology of Polysynthesis. There’s a lot of different ways of making complex verbs, and this is just one of them.

Thank you! This looks like just what I was looking for.

Is -yos(y)- suppletive, or related to -is- somehow? It doesn't seem to match the other A/B versions in some way, although the patterns are rather disparate.

Yes, the protoforms are *yusaa- vs *yusaye-. Syncope works in weird ways. (See also: Old Irish.)

I forgot about Eŋes having syncope. That explains it.

More broadly, nested SVCs are reasonably common in general, both in Eŋes and in some natlangs. Even if this particular construction is rarely used, it’s a good introduction to the type.

(It’s not like this is made up, even. My perfectionism meant that I actually researched birds of prey to work out what a sensible object for that sentence would be. Apparently lizards are a major component of the diet of the wedge-tailed eagle, for instance.)

Don't worry, my perfectionism would have led to creating (incorrect) beliefs about what Eŋes speakers think birds of prey eat, and then researching such beliefs in a late night attack of the ‘is it indeed realistic that Eŋes speakers think that birds of prey might eat X’. The research load is additionally one reason why my sentient corvid language is rather stuck in the ‘massive reworking as I finally understand what a choana is’ stage.

[bradrn]

I really like how the Eŋes verb complex is shaping up. It reminds me a bit of how I'd like Lēri Ziwi's verbs to be – just about that level of complexity, although marking different things differently.

On this point, I can recommend Matisoff’s Typology of Polysynthesis. There’s a lot of different ways of making complex verbs, and this is just one of them.

Thank you! This looks like just what I was looking for.

You’re welcome! I believe it was originally Vardelm (no longer very active here) who pointed me to it. It’s only a starting point, of course, but a good one.

(It’s not like this is made up, even. My perfectionism meant that I actually researched birds of prey to work out what a sensible object for that sentence would be. Apparently lizards are a major component of the diet of the wedge-tailed eagle, for instance.)

Don't worry, my perfectionism would have led to creating (incorrect) beliefs about what Eŋes speakers think birds of prey eat, and then researching such beliefs in a late night attack of the ‘is it indeed realistic that Eŋes speakers think that birds of prey might eat X’. The research load is additionally one reason why my sentient corvid language is rather stuck in the ‘massive reworking as I finally understand what a choana is’ stage.

Never fear, I’m not even at that state yet… though, that said, premodern peoples tend to have reasonably accurate ideas about animal behaviour when they’re able to see it.

(What is a choana, anyway? And why does it matter?)

[Lērisama]

Never fear, I’m not even at that state yet… though, that said, premodern peoples tend to have reasonably accurate ideas about animal behaviour when they’re able to see it.

They're weird foreign mostly-mythologised birds¹ then. Sorted.

(What is a choana, anyway? And why does it matter?)

It's a hole in the mouth where in humans the velum is, linking directly into the nasal cavity. I was under the mistaken² impression that it could be closed, allowing non-nasal anything, and plosives of any kind, but it can't be, so everything is always nasalised and plosives are impossible.

¹ I.e. the speaking ones I'm trying to make a language for.
² I think. Most information about avian anatomy is not the most focused on vocalisation, except for the syrinx, where the problem is working out which of the masses of information is actually relevant.

[Glenn]

Thank you for continuing to share this, especially the description of motional SVCs. I have found some aspects of Eŋes extremely difficult to get my head around, but I found the motional SVCs easier to understand.

[bradrn]

Thank you for continuing to share this, especially the description of motional SVCs.

You’re welcome! It makes it worth it when people are interested.

I have found some aspects of Eŋes extremely difficult to get my head around

Hmm… like what exactly? I’m happy to write more about anything unclear.

[Glenn]

I have found some aspects of Eŋes extremely difficult to get my head around

Hmm… like what exactly? I’m happy to write more about anything unclear.

I'm afraid that I would have to spend some time re-reading the previous posts on Eŋes to be able to answer that question properly. I do recall having some difficulty seeing how the surface forms of some words were produced from their roots, or teasing out the structure of words with more complicated morphology.

[bradrn]

I have found some aspects of Eŋes extremely difficult to get my head around

Hmm… like what exactly? I’m happy to write more about anything unclear.

I'm afraid that I would have to spend some time re-reading the previous posts on Eŋes to be able to answer that question properly. I do recall having some difficulty seeing how the surface forms of some words were produced from their roots, or teasing out the structure of words with more complicated morphology.

Hmm, OK. I felt the morphophonology wasn’t particularly complex, but please tell me if you can point to any specific examples which confuse you!

[Glenn]

Hmm, OK. I felt the morphophonology wasn’t particularly complex, but please tell me if you can point to any specific examples which confuse you!

I suspect that it was my own lack of understanding, as well as feeling intimidated by some of the unglossed examples. When I have time, I will re-read the previous material on Eŋes, and if I have any particular questions, I will let you know.

[bradrn]

Hmm, OK. I felt the morphophonology wasn’t particularly complex, but please tell me if you can point to any specific examples which confuse you!

I suspect that it was my own lack of understanding, as well as feeling intimidated by some of the unglossed examples. When I have time, I will re-read the previous material on Eŋes, and if I have any particular questions, I will let you know.

Thanks!

[bradrn]

Interlude: A protolanguage

This is, in fact, an update of one of my very first conlangs, all the way back in 2015. But I happened to rediscover it recently, and it occurred to me that it would make an excellent protolanguage for not just one, but two of my latest conlang drafts. Naturally, I’m taking the opportunity to revise it quite extensively, given my now much improved knowledge of linguistics. That said, for now it’s just a fairly simplistic set of ideas which I wanted to get out of my head so I could concentrate on other stuff…

Phonology

The consonant inventory in onsets is relatively straightforward, as follows:

m n
p t k
b d g
f s š h
v z ž
l (y)

You will note that I don’t use IPA here. That’s because this is a reconstruction and I can’t say for sure what any of the phonetic values were. Some notes, though:

• The exact place of articulation of š ž is uncertain, beyond being postalveolar. Some descendants reflect them as a rhotic, suggesting that they may have been retroflex; others reflect them as palatal. (I tend towards reconstructing retroflexed [ʂ ʐ].)
• It seems very likely that h was pharyngeal [ħ] in at least some environments. Though most descendants have no pharyngeals, some retain [ħ] as an allophone, or even as an independent phoneme (though there is no evidence to suggest reconstructing [ħ] as a separate phoneme in the protolanguage).
• It seems clear that the protolanguage had one liquid or rhotic consonant, but its identity is uncertain. I write l here, but could as easily have written r.
• The phonemic status of y is doubtful: it was probably an allophones of i (as also suggested by the vowel system, see below). Here I write it with y next to vowels, and as i elsewhere.

The vowel system is a more difficult matter. It’s easier to list the syllable rimes, which appear to have been organised in three grades:

a e i o
an en in on
ay ey iy oy

The phonetic identity of these grades is less certain:

• The plain grade a e i o probably had something more or less like their IPA values. The last vowel could probably be written u as easily, but its basic value was probably [o] or even [ɔ] — occurrences of /o/ in descendents tend to be cognate, whereas occurrences of /u/ aren’t, suggesting that /u/ is a secondary development in those languages which have it.
• The nasal grade an en in on had some variety of nasalisation on top of the basic vowel. In descendants the nasal component appears to have interacted with the following consonant as much as with the vowel, so I typically write it as one of the nasals ⟨m n⟩ depending on the next phoneme (or as ⟨ɴ⟩ in cases of ambiguity). Alternately, the vowels themselves may have been nasalised, as found in several descendants.
• The palatal grade ay ey iy oy is least certain. In some descendants they have the effect of palatalising the following consonant, which suggests the diphthongal transcription used here — though that does raise the question of why i is non-palatalising. Quite probably there was some form of vowel quality difference too. No descendant maintains this grade as fully separate from the other two.

These three grades become rather important in the morphology — many suffixes select a particular grade for a preceding vowel. Importantly, grades may not be combined: thus if a nasal vowel (say) occurs before a palatal suffix, the result is palatal, not nasal+palatal.

Furthermore, in some environments the vowel grades are non-contrastive. Before a nasal consonant, vowels always reflect the nasal grade. Similarly, before i/y, vowels always reflect the palatal grade.

Each syllable consists of at least one of the above syllable rimes, optionally preceded by a consonant. Whether codas are allowed depends on how the rimes are analysed, but even under the maximal analysis they would have been restricted to [m n j]. Consonant clusters are forbidden (again modulo the analysis of vowel grades), but hiatus is freely allowed. It seems very unlikely that sequences of vowel.nasal+vowel in hiatus were regularly distinguished from vowel+n+vowel, or that vowel.palatal+vowel were distinguished from vowel+i+vowel — at least on a surface level (since underlyingly there are certainly differences between roots ending in a nasal-grade vowel and roots ending in a nasal consonant).

Nouns

Nouns can be reconstructed as having been marked for eight cases (nominative, accusative, genitive, instrumental, dative, locative, ablative, comparative) and two numbers (singular, plural), split across two genders (masculine, feminine).

A word on the genders first. There is a pervasive opposition in both declension and conjugation between consonant-final roots and vowel-final roots (or, if you really must, ‘thematic’ and ‘athematic’). In the nominal domain, these have become strongly associated with gender: consonant-final roots are almost all masculine, and vowel-final roots are almost all feminine. The exceptions are almost all words where semantic gender overrides the phonological gender assignment, e.g. aheɴ- ‘man’ is vowel-final, yet masculine.

In the singular, there are two declensions, one for each gender, as follows:

Case	Masculine	Feminine
Nominative	nola ‘hand’	žan ‘woman’
Accusative	nolali	žanli
Genitive	nolade	žande
Instrumental	noley	žaney
Dative	nolen	žanhen
Locative	nolive	žayve
Ablative	nolema	žama
Similative	noleko	žanko

In the plural, there are again two declensions, but only the plural marker changes with gender — the case ending is always the ‘feminine’ / V-final form:

Case	Masculine	Feminine
Nominative	nolin ‘hands’	žanya ‘women’
Accusative	nolinli	žanyali
Genitive	nolinde	žanyade
Instrumental	noliney	žanyaey
Dative	nolinhen	žanyahen
Locative	nolimve	žanyayve
Ablative	nolima	žanyama
Similative	nolinko	žanyako

Adjectives inflect for case and number like nouns (agreeing with their head within the NP). They also agree for gender, taking masculine or feminine endings as appropriate.

Pronouns allow an additional number, namely the dual. First- and second-person pronouns can be reconstructed as follows:

	SG	DU	PL
1	na	nagon	zan
2	ye	yegon	yen

Third-person pronouns do not allow the dual, but vary for gender:

	SG	PL
M	haz-	hazin
F	le	lenya

All these pronouns are marked for case using the same suffixes as nouns. The ‘feminine’ suffixes are used (nali, nade, etc.), with the exception of the 3sm pronoun (haza, hazali, etc.).

Verbs

Verbs can be reconstructed as having been marked for five versions (stative, dynamic, transitive, benefactive, malefactive), four tenses (remote past, past, present, future), five moods (indicative, imperative, interrogative, conditional, optative), eight person–gender–number values, and two polarities.

The verb root generally has a CVC- shape, for instance žog- ‘eat’. Some verbs have more syllables while still ending in a consonant, e.g. eniz- ‘think’. A few verbs end in a vowel, e.g. ka- ‘to give’.

The first suffix on the verb is a somewhat fused marker for mood and what I call ‘version’ — basically a sort-of-classifier encompassing notions of transitivity, volition and Aktionsart — as follows:

	Stative	Dynamic	Transitive	Benefactive	Malefactive
Indicative	-a-	-e-	-o-	-yahe-	-ida-
Imperative	—	-en-	-on-	-yahen-	-ina-
Interrogative	-aza-	-aze-	-azo-	-ažahe-	-azida-
Conditional	-ya-	-ye-	-yo-	-yiye-	-yda-
Optative	-ea-	-ea-	-oy-	-eyhe-	-eda-

(Note that the conditional and benefactive forms change the grade of the last vowel irrespective of any intervening consonant: thus žog-yo- becomes žoygo-, and eniz-yahe- becomes eniyzahe-.)

[I pause here for an external note: the version category was called ‘voice’ in my 2015 grammar, because at the time I had no real idea what grammatical voice actually was. But I like the basic idea, so I revived it based on my vague memory of similar ‘theme’ markers in Athabaskan, plus seeing Man in Space’s recent post on Kgáweq’, plus my vague discontent with how zompist did benefactivity in Kebreni.]

The tenses are marked as follows: remote past -ve, past -de, future -go, and present unmarked (as is so common). Not all tenses are compatible with all moods: the imperative is limited to the present, and the optative is limited to the future.

The PGN markers are fairly transparently derived from the pronouns:

	SG	PL
1	-n	-zan
2	-y	-yen
3M	-za	-zin
3F	-le	-len

Finally, a negative marker -ve can come at the end.

Some basic examples of the system thus far in action:

Abo enizale. The person is thinking.
Abo nali eniyzahele. The person is thinking of me.
Žoygahenzan. Let’s eat.
Žogeagoyve. May you not be eaten.
Hazali gošehen abohen kaydagomve. I would not give it to the big person [because it would disadvantage them].


To be continued… sometime. Once I’m done with the next few Eŋes posts, at least.

[WeepingElf]

Looks nice. I like it.

[Travis B.]

I have a suggestion ─ have i be realized as [ɨ] but iy be realized as [iː], explaining why the former is non-palatalizing while the latter is palatalizing.

[Lērisama]

I giev you warm fuzzy feelings and a question: in the dative, why does the thematic¹ singular dative have -en in nolen, but -hen in the plural nolinhen and zalinhen. Is the former supposed to be nolahen, or is it just a bit irregular?

¹ I really must; I've just been reading a lot of IE stuff, and I'm in quite a Hittitey mood right now

[bradrn]

I have a suggestion ─ have i be realized as [ɨ] but iy be realized as [iː], explaining why the former is non-palatalizing while the latter is palatalizing.

Could be. It’s a protolang, remember; feel free to choose your own phonetics!

(Personally, I had a similar thought, but with [ɪ].)

I giev you warm fuzzy feelings and a question: in the dative, why does the thematic¹ singular dative have -en in nolen, but -hen in the plural nolinhen and zalinhen. Is the former supposed to be nolahen, or is it just a bit irregular?

Just a bit irregular. Note that there is no single ‘thematic vowel’, like there is in IE.

EDIT: whoops, I misread the question! It’s because the plural suffixes are vowel-final and hence always take the ‘athematic’ suffixes.

Also, do keep in mind that much of this was written at 1am and may be somewhat half-baked…

[Lērisama]

Note that there is no single ‘thematic vowel’, like there is in IE.

I agree that vowel/consonant-final is better; calling it thematic was just some combination of a bad attempt at humour and what happens at 11pm when you have just been reading about IE verbs.

EDIT: whoops, I misread the question! It’s because the plural suffixes are vowel-final and hence always take the ‘athematic’ suffixes.

Oh, that was my misreading, I had nol-a as vowel final and žan as consonant final because I forgot about the nasal grade¹

Also, do keep in mind that much of this was written at 1am and may be somewhat half-baked…

I didn't notice; it looks very well done

¹ I didn't make that mistake in the plurals somehow. Brains are weird)