Aseryla

Intermediate language is a sublanguage that can represent each grammatical relation found after do the sentences analysis.

It only has the objective of facilitate the translation from a constituent tree to internal language (the format that the "memory of this system" can manage).

Noun Relations : NP/subject(nouns)/modifier/{ modifier /...}{*NP...}

Modifier = (tendency)(type)(term)/
Tendency = + (positive ), - (negative)

Type = J (adjective), V (verb) [if the type is noun, then is added to the main concept: specializations (_ is the whitespace separator)]

(NP (DT a) (ADJP (JJ nice) (NN dog))

→ NP/dog/+Jnice

(NP (DT the) (ADJP (JJ red) (CC and) (RB not) (JJ blue) (VB jumping) (NN man) (NN cat))

→ NP/man_cat/+Jred/-Jblue/+Vjumping

"man_cat" is an specialization

(NP (DT the) (NN cat))

→ NOTHING (the object "cat" will be managed from the syntax structure)

* is for separate the possible multiple subjects in the noun phrase;

only if personal prepositions appear. E.g. "We, the metallic machine" → NP/We/*NP/machine/+Jmetalic/*

Verb Relations : V[P(ositive)|N(egative)]/verb(be,have,action)/[subject(noun)|-]{/object(noun){/object/...}}

Jhon had legs

→ VP/have/Jhon/legs

Tigers run fast

→ VP/run/Tigers

dogs and cats have not been running in the garden nor the beach

→ VN/running/dogs/garden/beach

→ VN/running/cats/garden/beach

a pen or an stapler can be thrown through the window

→ VN/thrown/-/pen/stapler

if the subject is the symbol '-', means that no subject has relation with the indicated verb

Attribute associations : WT/subject(noun)/mainPP(noun)/{modifier/{modifier/...}}{*WT...}

man with hands

→ WT/man/hands

"cat" (subject) "with big claws and large legs"

(NP (NP (NNS cats)) (PP (IN with) (NP (NP (JJ big) (NNS claws)) (CC and) (NP (JJ large) (NNS legs)))

→ WT/cat/claws/+Jbig/*WT/cat/legs/+Jlarge/*

Numeral associations : CN/noun/number

"I have 3.5 legs and four hands"

...

num(legs,3.5) → CN/legs/3 (integer part)

num(hands,four) → CN/hands/4 (normalizedNER)

...

"There is two(400.0) hundred(400.0) people"

number(four, hundred) [it's no processed]

num(people, hundred) → CN/people/400

"moved by one(>=1) or more things"

normalizedNER → >=1 !! (it removes everything that is not a number)

CN/things/>=1 → CN/things/1

* num and number are indicators that appears the <dependencies> of the xml output results file obtained from Stanford NLP core.

Specializations: SP/noun_noun{_noun....}

The symbol _ is the whitespace separator

"Ring Storm is a heavy metal band"

→ SP/Ring_Storm

→ SP/heavy_metal_band

There will be translated by the following internal language WORD codes:

→ WORD N Ring ring 1

→ WORD N Storm storm 1

→ WORD N heavy heavy 1

→ WORD N metal metal 1

→ WORD N band band 1

→ WORD N Ring_Storm ring_storm 1

→ WORD N heavy_metal_band heavy_metal_band 1

Proper Nouns: PN/type/noun_noun{_noun....}

Type = 1 (organization), 2 (location) or 3 (person)

The symbol _ is the whitespace separator

"John Miller was born in New York City and works for ABC"

→ PN/2/Jhon_Miller

→ PN/3/New_York_City

→ PN/1/ABC

There will be translated by the following internal language WORD codes:

→ WORD N John person 1

→ WORD N Miller person 1

→ WORD N New location 1 (lemmatization problem)

→ WORD N York location 1

→ WORD N City location 1

→ WORD N Jhon_Miller person 1 (note the lemma is not person_person)

→ WORD N New_York_City location 1

→ WORD N ABC organization 1

Intermediate Language