"Abrams wiped the table clean" Spanish translation options

The Spanish version of the MRS test suite has three sentences that are in English for some reason. I suspect there was some specific reason to not deal with them yet however I guess at this point I will just attempt to replace them with a Spanish version.

These are:

  • The tobacco garden dog barked.
  • Abrams wiped the table clean.
  • How happy was Abrams?

The first and the third are easy to translate keeping more or less the same syntax (well, the adjective structure in Spanish is realized via the particle de but otherwise it is the same recursive thing, I think).

The second sentence I am not sure. My L1 consultant suggests the following options:

  • Montse limpió correctamente la mesa.
    Montse cleaned well the table

  • Montse dejó la mesa limpia.
    Montse left the table clean

The consultant couldn’t come up with a direct translation of the English sentence, so that something like “wipe” and something like “clean” would both be present.

It therefore seems to me that the second option is what we want syntactically, so I am inclined to use that.


1 Like

wiped the table clean is a resultative construction. I understand the point of the MRS testsuite to be to illustrate different semantic constructions (and maybe to a lesser extent different syntactic paraphrases that lead to the same semantics). Does Spanish have resultative constructions?

Yes, I did my best to explain that to the consultant, and he basically says: well, the way we would say that is dejó la mesa limpia, so, “left the table clean”. He said, this implies the action of cleaning the table such that the result is that it’s clean. I am not sure about it though. Maybe best to look for a completely different sentence (which doesn’t involve table cleaning).

A quick look into the literature suggest Spanish might not have resultative constructions in the same sense English does. That would also explain why this particular item was left untranslated in the Spanish MRS test suite.

Now I am worried about the other two, which seemed easy for me :). One is adjective recursion (I am again thinking about it syntactically; adjective scope perhaps?) and the other is “How happy is Abrams?”

…I suppose I will just leave alone the items which were originally left in English by Montse. Clearly she did that for a reason (not only the resultative one but also the adjective one and the how happy… one).

Do you mean this example?

the tobacco garden dog barked.

If so, I think what this is an example of is ambiguous noun compounding, not anything adjective related. Both [[tobacco garden] dog] and [tobacco [garden dog]] are licensed in English. I’m no expert in Spanish, but I suspect that this construction would be translated using prepositions (like in other Romance languages), so maybe that ambiguity isn’t present?

1 Like

That may well be the case, TJ, thank you for this comment! You are correct about Spanish in this case, it would be el perro del jardin de tobacco and I think no ambiguity is present.

1 Like

Thank you, @olzama, for creating this discussion! Translating the MRS test suite for Portuguese is another task that I am also interested in. In 2020, I created a list with some sentences from the MRS test suite on the Tatoeba website:


I just added the 3 sentences with my translations for Portuguese:

  1. O cachorro da tabacaria latiu.
  2. Abrams deixou a mesa limpa.
  3. Quão feliz estava Abrams?

In (1), the world “tabacaria” (A place to buy tobacoo and related stuff, OpenWordnet-PT Browser 10713923-n) can be the translation from “tobacco garden”? Google found a website for a place called “Tobacco Garden Bar”. Thinking about what @trimblet said. I can see one alternative without any ambiguity.

1’. O cachorro do jardim de tabaco latiu.

The article [1] presents the sentence (2) in the example 16.b but didn’t propose a translation. It is presented as one example of “weak resultative meaning”:

weak resultatives are those in which the result predicate denotes either the purpose or the conventional result of the activity encoded in the main verb

In Portuguese (and Spanish) the verb (deixar/left) is not semantically related to the adjective (limpa/clean). I didn’t read the all the article, but the translation does not preserve the semantic connection between verb and adjective. I am not sure why the authors conclude with

True resultatives,which involve both Manner and Result, are attested in English, but not in EP.

On the other hand, ERG trunk version licensed two readings for the English version of (2). The first one is strange to me because the verb complement is not a ‘table that is clean’ but the ‘table’ ended up ‘clean’ because of the event ‘wipe’. Same analysis for the translation, so maybe the second analysis should be the golden analysis? But it is the first analysis the one recorded in erg/tsdb/gold/mrs, right @Dan ?

[ LTOP: h0
INDEX: e2 [ e SF: prop TENSE: past MOOD: indicative PROG: - PERF: - ]
RELS: < [ proper_q<0:6> LBL: h4 ARG0: x3 [ x PERS: 3 NUM: sg IND: + ] RSTR: h5 BODY: h6 ]
 [ named<0:6> LBL: h7 CARG: "Abrams" ARG0: x3 ]
 [ _wipe_v_1<7:12> LBL: h1 ARG0: e2 ARG1: x3 ARG2: x9 [ x PERS: 3 NUM: sg IND: + ] ARG3: h10 ]
 [ _the_q<13:16> LBL: h11 ARG0: x9 RSTR: h12 BODY: h13 ]
 [ _table_n_1<17:22> LBL: h14 ARG0: x9 ]
 [ _clean_a_of<23:28> LBL: h15 ARG0: e16 [ e SF: prop TENSE: untensed MOOD: indicative ] ARG1: x9 ARG2: u17 ] >
HCONS: < h0 qeq h1 h5 qeq h7 h10 qeq h15 h12 qeq h14 >
ICONS: < > ]

[ LTOP: h0
INDEX: e2 [ e SF: prop TENSE: past MOOD: indicative PROG: - PERF: - ]
RELS: < [ proper_q<0:6> LBL: h4 ARG0: x3 [ x PERS: 3 NUM: sg IND: + ] RSTR: h5 BODY: h6 ]
 [ named<0:6> LBL: h7 CARG: "Abrams" ARG0: x3 ]
 [ _wipe_v_1<7:12> LBL: h9 ARG0: e2 ARG1: x3 ARG2: x10 [ x PERS: 3 NUM: sg IND: + ] ]
 [ _the_q<13:16> LBL: h11 ARG0: x10 RSTR: h12 BODY: h13 ]
 [ _table_n_1<17:22> LBL: h14 ARG0: x10 ]
 [ subord<23:29> LBL: h1 ARG0: e15 [ e SF: prop ] ARG1: h16 ARG2: h17 ]
 [ _clean_a_of<23:28> LBL: h18 ARG0: e19 [ e SF: prop TENSE: untensed MOOD: indicative ] ARG1: i20 ARG2: u21 ] >
HCONS: < h0 qeq h1 h5 qeq h7 h12 qeq h14 h16 qeq h9 h17 qeq h18 >
ICONS: < > ]

[1] https://revistes.uab.cat/isogloss/article/view/v8-n5-yao

1 Like

Ah, but it is my understanding is that ambiguity is what we want here? In other words, the purpose of this item in the MRS test suite is to produce an ambiguous structure?

1 Like

In Portuguese we can have a syntactic ambiguity but it would be a crazy semantics to say

A dog of tobacco from the garden

1 Like

I asked my Puerto Rican (linguist) co-worker about this sentence and he said the same thing :slight_smile:

1 Like

No, @olzama, the intent of the MRS test suite is not to present items that are ambiguous (though this cannot always be easily avoided), but instead to present items that each illustrate some specific syntactic phenomenon and its associated MRS (for example, a bare plural or a compound noun).

1 Like

The gold MRS for “Abrams wiped the table clean” is intended to express the resultative meaning, @arademaker, but I can see the notion of a resulting state is not easy to read in the MRS. Note the subtle contrast of this MRS with the MRSs for “Abrams wiped the clean table” or “Abrams wiped the table which was clean”, where in both, the labels of the EPs for “table” and “clean” are identified, different from the resultative, where the labels are distinct. This MRS for the resultative is also consistent with the (intended) MRS for “Abrams caused the table to be clean”.


The predicate-argument structure is always more salient to me. So far, labels are mainly names for scopes for me, but it seems that there is something more than that. In other words, how the translation to FOL may be different from an adjective and noun sharing a labels vs not sharing the same label?

@Dan , yes, but I meant that particular item, the tobacco garden dog barked. TJ mentioned the ambiguity there.