Skip to content

Instantly share code, notes, and snippets.

@addrummond
Last active December 16, 2015 10:39
Show Gist options
  • Save addrummond/5421575 to your computer and use it in GitHub Desktop.
Save addrummond/5421575 to your computer and use it in GitHub Desktop.
This is a Python script for computing the strong paradigm for Icelandic 'kaldur' ('cold') given the toy analysis presented in Drummond & Hornstein (forthcoming).
# coding=utf-8
import sys
from itertools import *
from sets import *
vocab_items = [(x[0], Set(x[1])) for x in
(u'kalt', ()),
(u'kaldur', ('+A', '+B')),
(u'köld', ('-B',)),
(u'kaldan', ('-pl', '+acc', '+A', '+B')),
(u'kalda', ('-pl', '+acc', '+A', '-B')),
(u'köldum', ('+dat', '+A', '+B')),
(u'kaldri', ('-pl', '+dat', '-B')),
(u'köldu', ('+dat', '-A', '+B')),
(u'kaldrar', ('-pl', '+gen', '-B')),
(u'kalds', ('-pl', '+gen', '+A', '+B')),
(u'kaldir', ('+pl', '+A', '+B')),
(u'kaldar', ('+pl', '+A', '-B')),
(u'kaldra', ('+pl', '+gen', '+A', '-B')),
]
impoverishment_rules = [(Set(x[0]), Set(x[1])) for x in
(('+gen',), ('+A',)),
(('+pl', '+gen'), ('-B',)),
(('+pl', '-A'), ('-B',)),
(('+pl', '+dat'), ('+A', '+B', '-pl')),
(('+pl', '+acc', '+B'), ('-pl', '-B'))
]
def proper_subset(specification1, specification2):
left = False
right = False
for x in specification1:
if x not in specification2:
left = True
break
for x in specification2:
if x not in specification1:
right = True
break
return right and not left
def check_match(specification, item):
for f in specification:
if f not in item:
return False
return True
def apply_impoverishment_rule(rule, item_specification):
item_specification = [x for x in item_specification]
applied = False
if check_match(rule[0], item_specification):
applied = True
for f in rule[1]:
ff = (f[0] == '+' and '-' or '+') + f[1:]
if ff in item_specification:
i = item_specification.index(ff)
item_specification[i] = f
elif f not in item_specification:
item_specification.add(f)
return tuple(item_specification), applied
def filter_by_order(vocab_items):
lattice_levels = [[vocab_items[0]]]
for it in islice(vocab_items, 1, None):
inserted = False
for j in xrange(len(lattice_levels)):
if proper_subset(it[1], lattice_levels[j][0]):
inserted = True
if j == 0:
lattice_levels.insert(0, [it])
else:
lattice_levels[j-1].append(it)
break
if not inserted:
for it2 in lattice_levels[-1]:
if proper_subset(it2[1], it[1]):
inserted = True
lattice_levels.append([it])
break
if not inserted:
lattice_levels[-1].append(it)
return lattice_levels[-1]
def execute(impoverishment_rules, vocab_items, item_specification):
for r in impoverishment_rules:
item_specification, applied = apply_impoverishment_rule(r, item_specification)
matching = filter_by_order([x for x in vocab_items if check_match(x[1], item_specification)])
if len(matching) > 1:
print "\n\nAmbiguity:", matching
return None, False
if len(matching) == 0:
return None, False
return matching[0], True
def print_icelandic_table():
for pl in [['-pl'], ['+pl']]:
print (pl[0][0] == '-' and "[Singular]" or "[Plural] ") + " MASC FEM NEUTER"
for case in [['+nom'], ['+acc'], ['+dat'], ['+gen']]:
sys.stdout.write(case[0][1:].upper() + (" " * (11-len(case[0][1:]))))
for gender in [('masc', ['+A', '+B']), ('fem', ['+A', '-B']), ('neuter', ['-A', '+B'])]:
fs = [x for x in pl]
fs.extend(gender[1])
fs.extend(case)
fs = tuple(fs)
x, ok = execute(impoverishment_rules, vocab_items, fs)
if not ok:
return
sys.stdout.write(x[0])
if gender[0] != 'neuter':
sys.stdout.write(" " * (11-len(x[0])))
else:
print
print
print_icelandic_table()
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment