General proposal for spatial deixis

[rueter]

Hi, @nschneid and @Stormur!
Thanks for your feedback!
Looking at the terminology offered at https://glossary.sil.org/term/boundedness, it appears that spatial deixis might offer one possible solution with either the feature Boundedness=Yes on the adverbs “out”, “in”, “up”, below OR a relation advmod:bounded .

The approach would probably be to only use advmod:bounded where there is a distinction.
Definition: Bounded spatial deixis is place deixis that has a component of meaning indicative of a border.
The examples are out there, in here vs there, here:
advmod:bounded(there, out)
advmod:bounded(here, in)

The example sentence from https://universaldependencies.org/u/dep/all.html#al-u-dep/case the cafe up beside the lookout indicates a structure:
det(cafe, the)
nmod(cafe, lookout)
case(lookout, up)
case(lookout, beside)
det(lookout, the)

Instead of using the relation case(lookout, up), which might imply that the word “up” is some kind of ADP, we could decide that the ADV “up” has a hypothetical relation advmod:bounded(lookout, up).

The fact that the position taken by the ADV “up” preceding an ADP can be occupied by other words, e.g., “out”, “in”, “down”, “over”, which with other ADPs: “back behind the barn”, would seem to fit the description. This makes us wonder:

• do other languages might have this same kind of construction?
• is this construction related in any way to English (Salish Sea): “down south”, “back east”, “up north”, “out west”. The expressions “down to Vancouver (Vancouver, Washington)”, “up to Vancouver (Vancouver, B.C.)” “over to Spokane (Spokane, Washington, on the east side of the Cascade mountain range)”, but “over/out/across to Bremerton (Bremerton is located across Puget Sound from Kent, Washington)” also indicate relative points of the compass.
• My gut feeling is that these are not the same relations as expressed by particle verbs of the type:
  • “hike up to Moose Cow Lake” ‘implies the hike was begun at a point south of or lower in altitude to the lake’
  • “hike in to Moose Cow Lake” ‘implies the hike goes into/through a forest’
  • “hike down to Moose Cow Lake” ‘implies the hike was begun at a point north of or higher in altitude to the lake’
  • “hike to Moose Cow Lake” ‘neutral end point is lake’

My working knowledge of Finnish says that ego-related indication of altitude is not used to express north or south. Instead, ylhäällä vuoren huipulla on kahvila. ‘up on top of the mountain is a cafe.’

Here we could interpret this as:

• “ylhäällä<Adv+Ade>” ‘up (above)’ is a general place, and “vuoren<mountain+Gen> huipulla<on.top+Ade>” ‘on top of the mountain’ further specifies the place. Perhaps, this could be an appos relation between spatial adverbs, or
• Unlike the English adverb “up”, the Finnish adverb “ylhäällä<Adv+Ade>” can stand alone, so the two do not represent the same relation.

The Erzya equivalent “вере, пандо прясо, кафе.” ‘up<Adv+Loc>, hill top<+Ine>, cafe.’
cop(вере, кафе)
compound(прясо, пандо)
appos(вере, прясо)
seems to favor the same kind of interpretation as the Finnish one.

@ftyers, @jonorthwash, @garanes, @jasiewert, @nikopartanen, @flammie, @amir-zeldes, @dan-zeman !

Do you know of other languages where the cluster “up beside the lookout” might be interpreted as an expression of bounded spatial deixis? Granted, what I describe above is the use of a feature for the ADV “up”, whereas I should be looking for a deprel advmod:bounded. Personally, I would like to avoid introducing any new subrelations.

Since the deprel advmod:bounded would seem to involve the same elements, it might be just as well to expand the definition of case. What do you think?

[dan-zeman]

In Czech, you could indeed say tam venku "there out", or tam nahoře "there up", as well as venku u zídky "out by the wall", nahoře u vyhlídky "up beside the lookout", tam nahoře u vyhlídky "there up beside the lookout" etc. I fail to see why anything would be better than simply treating them as multiple adverbial/place modifiers of the same head.

(NB: You could also use nahoře "up" as a metaphor for north, and dole "down" for south.)

[rueter]

In Czech, you could indeed say tam venku "there out", or tam nahoře "there up", as well as venku u zídky "out by the wall", nahoře u vyhlídky "up beside the lookout", tam nahoře u vyhlídky "there up beside the lookout" etc. I fail to see why anything would be better than simply treating them as multiple adverbial/place modifiers of the same head.
Thank you @dan-zeman,

You mention tam nahoře u vyhlídky "there up beside the lookout".
Is this then
appos(tam, vyhlídky)
case(vyhlídky, nahoře)
case(vyhlídky, u)
OR
appos(tam, vyhlídky)
appos(tam, nahoře)
case(vyhlídky, u)
OR...

(NB: You could also use nahoře "up" as a metaphor for north, and dole "down" for south.)
This is interesting, too. thanks!

[dan-zeman]

If the sentence is

Stalo se to tam nahoře u vyhlídky "It happened up there beside the lookout" (lit. happened REFL it there up beside lookout)

then I would do

advmod(Stalo, tam)
advmod(Stalo, nahoře)
obl(Stalo, vyhlídky)
case(vyhlídky, u)
expl:pv(Stalo, se)
nsubj(Stalo, to)

There would be no apposition.

[nschneid]

Consider "They live over [past the barn]." While "over" can be dropped, the PP "past the barn" cannot. I think this indicates that "over" modifies the PP. Because UD doesn't have ADP heads we have to say "over" modifies the noun "barn", but advmod because it is a modifier, not case. Only "past" would attach as case.

I believe this is consistent with @dan-zeman's Czech analysis.

[dan-zeman]

Define consistent :-)

It is different from my Czech analysis because I do not attach tam to u vyhlídky. But the situation in the language is also different because any of the three modifiers (tam, nahoře, u vyhlídky) can be dropped.