Clarification needed on relationship between entries and senses

[Dave]

In trying to create a LIFT export file, I am struggling to understand
the relationship between entries and senses. The specification reads,
"An Entry is made up of a number of senses. Each sense corresponds to
a part of speech." Does this imply that the part of speech used by a
sense must be unique across the entry? If so, what's the best way to
handle multiple definitions that happen to fall under the same part of
speech?

So, for example, for the English "run", which is appropriate?
(simplified format)

Entry: "Run"
- Sense 1: Noun, an act or instance, or a period of running
- Sense 2: Noun, a sequence of cards in a given suit
- Sense 3: Verb, to move with haste
- etc. (part of speech can be repeated)

OR

Entry 1: "Run"
- Sense: Noun, an act or instance, or a period of running
Entry 2: "Run"
- Sense: Noun, a sequence of cards in a given suit
Entry 3: "Run"
- Sense: Verb, to move with haste

The first certainly seems more logical, but doesn't seem to meet the
written requirements outlined in the specification about each sense
corresponding to a part of speech. Can you clarify?

Thanks,

Dave

[Ronald Moe]

As far as I know, the specification merely means that each sense can be
assigned to a different part of speech. So your first example is permitted.
However a lexicographer might choose to separate a lexeme form such as "run"
into separate entries so that one entry is for noun senses and the other
entry is for verb senses. Essentially this is treating the two entries as
homonyms:

Entry 1: "Run"
- Sense: Noun, an act or instance, or a period of running

- Sense: Noun, a sequence of cards in a given suit

Entry 2: "Run"

- Sense: Verb, to move with haste

- Sense: Verb, if the color in a piece of dyed cloth runs, the dye partly
dissolves and spreads

This is a matter of dictionary design. LIFT supports both (or should support
both). LIFT would also support your second example, although I have never
seen a dictionary that does this. I can imagine a situation in which a
lexicographer might do this for a meaning based dictionary. Such a structure
would facilitate sorting by semantic domain.

LIFT should not impose such structural decisions on the data. This kind of
decision should be left up to the lexicographer. Dictionaries tend to be
fairly similar in their structure due to the realities of language. A
standard like LIFT should enable a lexicographer to capture a variety of
structures within certain agreed upon limits.

Ron Moe

[John Hatton]

Hi Dave,

 

>"An Entry is made up of a number of senses. Each sense corresponds to a part of speech."

 

I'd say that wording qualifies as a bug in the spec (I’d remove the second sentence entirely).  You can report it here.

 

John Hatton

SIL PNG, Palaso, & SIL International Software Development

Google Talk chat: hattonjohn