Constructicon entry fields: a reference

Each construction is stored as an entry in the Italian Constructicon. Every entry contains a number of fields specifying formal, functional and usage information about the construction, as well as its relatedness to other entries. This page gives more detail on each field than the .yaml construction file walkthrough — it’s sourced from the project’s wiki.

ID

A unique and stable identifier for the constructional entry. This field is filled automatically as a new construction is created.

Name

Contains a “human-intelligible” name for the construction. To keep a consistent naming practice, we use the following decision tree to work out the name for each entry:

1. Is there a standard denomination for this construction (i.e., in the literature)? If so, use it. If there is no commonly accepted denomination, go to 2.
2. Can the name be expressed by means of its lexicalized components (e.g. “capo-N” for capo + Noun compounds)? If so, build the name that way. If there are no lexicalized components, go to 3.
3. Can the name be expressed by means of its syntactic relations (e.g. “Sbj V Obj1 Obj2” for the Ditransitive construction)? If so, build the name that way as long as it is easily readable. If it is not, go to 4.
4. Can the name be easily expressed by the function (e.g. “Contrastive Focus”)? If so, name the construction with reference to its function. If the name is too generic, or using the function as name is not possible for some other reason, go to 5.
5. If none of the previous criteria yield an easily readable/recognizable name, just be creative — we’ll find a better name with time!

CoNLL-X formalization

A reference to the construction’s .conllc file — see The CoNLL-C format.

Definition

A description of the construction’s function. Differently from the function metadata field in the CoNLL-C formalization (which should be as concise as possible), this is a non-formal, discursive description, similar to a lexicographic definition, and includes information on:

• semantics;
• pragmatic and discourse functions;
• information status and information structure features.

Differently from lexicographic definitions, it does not include examples (those go in a separate field, see Examples). Definitions can be more or less elaborate, depending on the information available in the literature. Example (from the salta fuori che V ‘turns out that’ entry):

A new piece of information comes to the speaker’s knowledge from an external source. The information acquired is often unexpected or contradicts the speaker’s expectations on the state of affairs, thus generating surprise in the speaker. However, since the moment of acquisition and the moment of enunciation are distinct, this construction does not convey that the speaker is currently surprised, but it is used to convey or generate surprise in the audience.

Restrictions

A discursive description of the constraints on the construction’s fillers. Such constraints can be hard (categorical) or soft (violable in specific cases), and can include:

• formally or semantically defined types/categories of lemmas and/or words (e.g. motion verbs, two-syllable words);
• specific lemmas/words;
• formal or semantic features of the fillers (e.g. a noun cannot be plural in a specific construction).

Constraints on the contextual and/or co-textual environment are not included here — see Usage and Notes instead. Restrictions can include both the ones already specified in the CoNLL-C formalization, and further constraints that aren’t/can’t be captured in the formalization.

Collocational preferences

A discursive field, similar to Restrictions, but for preferences rather than constraints — e.g. typical collocations for a slot. There’s currently no support for listing specific lemmas here; a description of the preferred filler types is used instead.

Usage

Up to 3 “usage” tags for the construction, used if it is constrained or shows preferences in its extra-linguistic behavior (i.e. if it’s typically or only found in some contexts). Three possible axes, one tag per axis:

• Style: Formal or Informal
• Modality: Spoken or Written
• Variety: Standard or Non-standard

This field cannot contain both tags of the same axis (e.g. both Spoken and Written) — in that case the construction isn’t marked for that axis at all, so the tag wouldn’t be informative. Further comments on usage (e.g. constraints on text genres) go in Notes. If none of the tags above is relevant, this field can be left empty.

Formal tags

Tags describing the construction from a formal point of view. Preferably drawn from the Comparative Concepts list of MoCCa (Model of Comparative Concepts for Aligning Constructicons), limited to the construction and strategy types. If no useful CC exists, other definitions shared in the (typological) literature can be used instead, e.g. prefixation, reduplication, impersonal construction.

Functional tags

Tags describing the construction from a functional point of view. Preferably drawn from the same Comparative Concepts list, limited to the meaning and information packaging types. If no useful CC exists, other definitions shared in the (typological) literature can be used instead.

Complexity level tags

The construction’s complexity level:

• word
• phrase
• clause
• beyond-clause

Category tags

One or more tags indicating the construction’s output category (ideally auto-filled from the .conllc file):

adjective, adverb, conjunction, discourse marker, interjection, noun, preposition, pronoun, verb, onomatopoeia, not applicable, other

Schematicity level

The construction’s schematicity level (ideally auto-filled from the .conllc file):

• lexically filled
• partially filled/schematic
• fully schematic

CEFR level

The language-competence level the construction belongs to: A1, A2, B1, B2, C1, C2.

Links

Two fields, ideally auto-filled from the .conllc file’s horizontal_links/vertical_links metadata (see CoNLL-C field reference):

• Horizontal links — constructions at the same level of abstraction that are semantically similar to/paraphrases of this one, or otherwise in a paradigmatic relation with it (e.g. antonymic constructions, or ones related by a derivational series).
• Vertical links — constructions at a higher level of abstraction; mainly instance links.

Examples

A collection of at least 5 examples from corpora, preferably from treebanks, with the construction annotated — see Annotating examples in CoNLL-Uc.

Notes

Free text for anything not covered elsewhere — e.g. if Usage is Non-standard, specify which variety (including diatopic), or the text genres it’s found in; frequency/emergence notes; etc.

References

Citations to the relevant literature on the construction.

Data collector

The name (or username) of whoever entered the construction.