Hermit Crab Parsing Engine Specification

5.7Phonetic Sequences and Phonetic Templates

5.7.1Purpose of Phonetic Sequences and Phonetic Templates

A phonetic template is a phonetic sequence augmented with a specification as to whether it must match against a phonetic sequence beginning at the left end of the phonetic sequence and/or ending at its right end. Phonetic templates are used internally and externally in the left environment, right environment, previous word, and next word fields of phonological rules.

The term “phonetic” is used here as a convenient cover term for sequences which are interpretable in phonetic terms. No theoretical stance is implied as to the existence of a distinction among morphophonemic, phonemic, or systematic phonetic levels.

The input to the morpher also represents a sequence of phonetic segments. However, as described in the chapter on the Preprocessor module, this input consists of a sequence of tokens, each containing a pair of strings (one of which represents the phonetic shape); there is no provision in the input to the morpher for the explicit representation of phonetic features. Likewise, the phonetic shape of a lexical entry is given as a string, not a phonetic sequence. However, the morpher uses phonetic sequences internally to represent an input word or its analysis. Finally, the Phonetic Output field of morphological rules allows affixal material to be specified by strings (to be translated into segments using a specified character definition table), as well as by phonetic sequences.

5.7.2Definitions

5.7.2.1Definition of Simple Context

Each Simple Context is a record data structure which defines a class of segments.

Record Label: simp_cntxt

Fields:

5.7.2.1.1Natural Class

Optionality: obligatory

Label: class

Type: atom (name of a natural class)

Purpose: To define the invariant features of the simple context. The natural class refers itself defines that feature content (see Natural Class—Features, section 5.6.2).

5.7.2.1.2Alpha variables

Optionality: optional

Label: alpha_vars

Type: alpha variable list

Purpose: To define the variable-valued features of the simple context. Variable-valued features are often used in assimilation rules; they are the “alpha” variables of classical generative phonology. Note that, unlike the standard notation for phonological rules, only the alpha variable and its polarity (+ or –) will appear here; the feature name to which the alpha variable refers is defined elsewhere in the rule (in the var_fs field).

5.7.2.2Segment

A segment data structure represents a single segment.

Record Label: seg

Fields:

5.7.2.2.1Representation

Optionality: obligatory

Label: rep

Type: string

Purpose: This gives the string representation of the segment, which should match the string representation of a segment in the appropriate character definition table.

5.7.2.2.2Character table

Optionality: obligatory

Label: ctable

Type: atom

Purpose: This gives the name of the character definition table in which the segment is defined.

5.7.2.3Boundary Marker

A boundary marker data structure represents a single boundary marker.

Record Label: bdry

Fields:

5.7.2.3.1Representation

Optionality: obligatory

Label: rep

Type: string

Purpose: This gives the string representation of the boundary marker, which should match the string representation of a boundary marker in the appropriate character definition table.

5.7.2.3.2Character table

Optionality: obligatory

Label: ctable

Type: atom

Purpose: This gives the name of the character definition table in which the boundary table is defined.

5.7.2.4Definition of an Optional Segment Sequence

An optional segment sequence is a record data structure which represents a sequence of one or more optional segments and/or boundary markers. Included in the data structure is an indication of how many times the optional segment sequence may appear within the string being matched.

Record Label: opt_seq

Fields:

5.7.2.4.1Minimum Occurrence

Optionality: optional

Label: min

Type: integer

Purpose: This defines the minimum number of times which the optional sequence may appear at the corresponding position in the string being matched.

If this field is omitted, the minimum number of times the optional sequence may appear is zero.

5.7.2.4.2Maximum Occurrence

Optionality: optional

Label: max

Type: integer

Purpose: This defines the maximum number of adjacent appearances of the optional sequence at the corresponding position in the string being matched.

If this field is omitted, the maximum number of times the optional sequence may appear is one.

Setting this field to -1 allows the optional sequence to match a stretch of segments any number of times. An optional sequence whose maximum occurrence is -1, and which contains an empty set of features (a set which matches any segment) will match any string. This will often be useful in morphological rules, where the entire length of a stem must be matched against, but only a small part of that stem need be specified in terms of its segmental content. An example is a prefix that attaches to any stem beginning with a consonant. The required phonetic input would be a single simple context for a consonant, followed by an optional sequence of any number of segments:

( )>)

(assuming the simple contexts “cons” and “mt” had been defined previously).

5.7.2.4.3Optional Sequence

Optionality: obligatory

Label: seq

Type: list

Contents: One or more names of simple contexts and/or boundary markers.

Purpose: This specifies the series of segments which are optional. The optional sequence is itself a phonetic sequence.

The simple context whose specification is the list (()) represents an optional sequence composed of any segment, and matches either a phonetic segment or a boundary marker. One use of such a sequence is described above under Maximum Occurrence.

5.7.2.5Definition of Phonetic Sequence

A Phonetic Sequence is a list representing a sequence of zero or more segments and/or boundary markers. Each member of the list is:

a Segment Record (defined above);

the name of a Simple Context (defined above); or

an Optional Segment Sequence (defined above).

Phonetic sequences are used in the input of morphological rules, the input and output of phonological rules, and in phonetic templates. However, Phonetic Sequences used for certain purposes must not contain Optional Segment Sequences; see section 5.7.2.6, Definition of Variable-Free Phonetic Sequences.

A phonetic sequence which is an empty list matches any single segment or boundary marker. These may be used in phonological rules of epenthesis or deletion, or in phonetic templates which represent variables in morphological rules, but would not normally appear in phonetic templates in the environment of phonological rules. (An empty environment can simply be omitted from a phonological rule.)

Example: The following phonetic sequence:

(

might represent a high vowel, followed by a morpheme boundary, followed by between zero and two open syllables.

5.7.2.6Definition of Variable-Free Phonetic Sequences

A Variable-Free Phonetic Sequence is defined as a Phonetic Sequence which does not contain any Optional Segment Sequences. Variable-free phonetic sequences are used internally to the morpher to represent sequences of segments in stems and words, and in certain parts of some rules.

5.7.2.7Definition of Phonetic Template

Functionally, a Phonetic Template is a phonetic sequence augmented with boundary conditions indicating whether it must match against another (variable-free) phonetic sequence beginning at the latter's left end and/or ending a the latter's right end. Structurally, a Phonetic Template is the record structure described below.

Phonetic Templates are used in the left environment, right environment, previous word, and next word fields of phonological rules.

Record Label: ptemp

Fields:

5.7.2.7.1Initial Boundary Condition

Optionality: optional

Label: init

Type: Boolean

Default: false

Purpose: If this field is true, the phonetic template must match against a phonetic sequence beginning with left-most segment of the latter, i.e. word-initially.

5.7.2.7.2Final Boundary Condition

Optionality: optional

Label: fin

Type: Boolean

Default: false

Purpose: If this field is true, the phonetic template must match against a phonetic sequence ending with the right-most segment of the latter, i.e. word-finally.

5.7.2.7.3Phonetic Sequence

Optionality: obligatory

Label: pseq

Type: phonetic sequence

Purpose: Represents the sequence of segments and/or boundary markers which must match against a phonetic sequence.

Directory: computing -> hermitcrab
computing -> Programme Specification for bsc Honours Computing, Graphics and Games
computing -> University of kent module specification template
computing -> Four box diagram Processor Output Input Main memory
computing -> Complete the following definitions with the words and phrases below
computing -> Geophysical Computing L02 Awk, Cut, Paste, and Join
computing -> Vce software Development: Programming requirements
computing -> Computing/Campus Network Services
computing -> Joint High Performance Computing Exchange (jhpce) Johns Hopkins School of Public Health
computing -> Office: fasb 267 Phone: 585-9792 Email
hermitcrab -> A new Program for doing Morphology: Hermit Crab

Download 403.76 Kb.

Share with your friends: