Attributes of a node

There are twelve attributes of a node at present.

 

Attribute Name Description
lemma lemma (see Lemma (lemma, SGML attribute <l>))
tag morphological category, tag (see Morphological marker (attribute tag, SGML attribute <t>))
form word-form after contingent adjustments (see Word-form (attribute form, SGML attribute <f>))
afun analytical function, or, to put it differently, type of relation to immediately governing node
lemid further/closer identification of lemma (esp. for lemmas which consist of more words)
mstag morphological-syntactic tag
origf original form of the word (see Word-form (attribute form, SGML attribute <f>))
origap formatting info preceding original word-form
gap1 formatting info preceding the word-form, part 1
gap2 formatting info preceding the word-form, part 2
gap3 formatting info preceding the word-form, part 3
ord serial number of the form in the sentence

From the point of view of manual annotation at analytical level only the attribute afun is essential, which is inserted in most of the cases manually. Attribute form represents the form of the word which figured as input in the morphological analysis and which almost always corresponds to the original form of the word in the text (for exceptions see Word-form (attribute form, SGML attribute <f>)). The attributes tag and lemma may contain values still undistinguished; this may happen in cases in which the manual annotation at morphological level failed to be carried out so far. In future the attribute lemid will be used to identify lexical units which consist of more words, or, as the case may be, to distinguish lexical units which behave in the same manner as far as morphology is concerned, but in a different way as regards syntax. The attribute mstag will be used as intermediate stage between morphological information (tag) and the so-called grammatemes at tectogrammatical level. It will, e.g., contain verbal tense, which cannot be stated at the morphological level (in full) - e.g., past tense.

The rest of the attributes is purely technical (origf, origap, gap1, gap2, gap3, ord). (In the programme GRAPH the value of the attribute origf appears at the lower margin of the screen.)

The attribute afun, which is the most important one from the point of view of the analytical level, does not relate directly to the node concerned, it belongs to the edge leading from the given node towards its governing (head) node. Only for technical reasons (since the necessity of assigning attributes to edges has been discarded) it is assigned to the subordinate node. In annotating, however, it is very useful to be aware of this fact since it helps to a better understanding of connections between some rules of annotation (e.g., thus the "meaningless" position of the function Pred, the special position of coordination and apposition, or the way of marking off a parenthesis become more clear and understandable).