Lingua::Phonology::Symbols 0.32

Lingua::Phonology::Symbols is a Perl module for associating symbols with segment prototypes.

SYNOPSIS

use Lingua::Phonology;
$phono = new Lingua::Phonology;

# Load the default features
$phono->features->loadfile;

# Load the default symbols
$symbols = $phono->symbols;
$symbols->loadfile;

# Make a test segment
$segment = $phono->segment;
$segment->labial(1);
$segment->voice(1);

# Find the symbol matching the segment
print $symbols->spell($segment); # Should print b

When using Lingua::Phonology, you usually manipulate Segment objects that have various feature values that specify the phonetic qualities of the segment. However, it is difficult to print those feature values, and a list of feature values can be difficult to interpret anyway. This is where Symbols comes in--it provides a way to take a Segment object and get a phonetic symbol representing the properties of that segment.

In Symbols, you may use add_symbol() to define text symbols that correlate to "prototypes", which are special Segment objects that represent the ideal segment for each symbol. After you have defined your symbols and prototypes, you may use spell() to find which prototype is the most similar to a segment in question, and get the symbol for that prototype.

As of v0.2, Symbols also includes diacritics. A diacritic is a special symbol that begins or ends with a *, and which is used to modify other symbols. If the best symbol match for a segment you are trying to spell is an imperfect match, Symbols will then attempt to use diacritics to indicate exactly how the segment is pronounced. For compatibility reasons, however, this feature is off by default. It can be turned on with set_diacritics.

You will probably want to read the add_symbol, spell, and loadfile sections, because these describe the most widely-used functions and the algorithm used to score potential matches. If youre not getting the results you expect, you probably need to examine the way your prototype definitions are interacting with that algorithm.