**SFUSED CANTONESE, Field Descriptions** John Alderete, Simon Fraser University ===== **EXAMPLE FIELDS** Example fields present the core facts as they are and do not typically require any analytical assumptions. However, they may include superficial facts about the error, like if the error is corrected, that are important to their analysis. **Example (exampleExample):** text, no value list, required. The longform of the error. A sentence or string of sentences containing at least one error, marked with a “/” prefix; may contain source words with a “^” prefix, and trigger words, shown with a “$” prefix; the context generally has at least ten characters before and after the error. **Intended Orthographic (exampleIntendedSpelled):** text, no value list, required. The orthographic form of the intended word or phrase. **Error Orthographic (exampleErrorSpelled):** text, no value list, required. The error word or phrase as it appears in the Example field. **Source Orthographic (exampleSourceSpelled):** text, no value list, required for contextual errors. The source word or phrase for the error as it appears in the Example field. **Intended Phonetic (exampleIntendedPhonetic):** text, no value list, required. The intended word or phrase transcribed using the phonetic transcription system described (essentially Jyutping, with some modifications). Error Phonetic (exampleErrorPhonetic): text, no value list, required. The error word or phrase transcribed using the phonetic transcription system. Source Phonetic (exampleSourcePhonetic): text, no value list, required for contextual errors. The source word or phrase for the error using the phonetic transcription system. **Wd Bounded Error? (exampleWordBounded):** text, YN value list, required. Specifies word-boundedness of sources in contextual errors: “Y” if the source is inside the error word, else “N”; compounds are considered single words, so errors with sources inside of them are word-bounded. **Corrected? (exampleCorrected):** text, YN value list, required. “Y” if the error is corrected by the speaker, “N” if not corrected. **Clipped Word? (exampleIncompleteError):** text, YN value list, required. “Y” means that the error word or phrase is incomplete and so it is suffixed with “=” in the Example field; “N” error term was completed **Notes (exampleNotes):** text, no value list, not required. Any special notes deemed important about the error, but not obviously covered in the other fields. Each note starts on a new line and generally starts with a note type that characterizes the note (these types include: Researcher, Observation, Assumption, Conjecture, Background, Issue). Some field values cannot be filled in because the value list does not allow a simple categorization and therefore have the value “See Notes”. In these cases, the notes provide specific information about these fields. These special notes are always prefaced with the field name it is further specifying. ===== **RECORD FIELDS** These are book-keeping facts about things like record ID numbers, the audio recording, the time stamp of the error, and the people involved (i.e., the speaker, the data collector, the analyst). **Record ID no. (recordErrorID):** number, no value list, required. The Record ID that gives a unique numerical value to every error; this field cannot be edited. **Total Completed (recordSumCompleted):** number, no value list, required. A book-keeping counter of the number of entries completed. **Last Modified (recordDateLastModified):** datestamp, no value list, required. Shows the date and time of the last edit to the record. **Record Completed? (recordComplete):** text, YN value list, required. Records start out with no value. “Y” = the record is complete and all necessary fields have been filled in. “N” = the record has an unresolved issue (used when database was in development). **Confirmed By (recordConfirmedBy):** text, Researcher value list, required. Labels for researcher or researchers who have verified the errors and provided the field values that classify the error. **Keep? (recordKeepYN):** text, YN value list, required. Many potential errors are submitted and later found not to be errors. “Y” = the record documents a true error (so it is kept), “N” = not a true error so not analyzed. **Reason to exclude (recordExclusionJustification):** text, no value list, required when Keep?=N. A brief explanation of why error was excluded, including things like belief that utterance reflects a habitual behavior or is a change of the speech plan. **Researcher (recordResearcherFound):** text, no value list, required. Label of the data collector, or collectors, that made the original submission. **Found Date (recordDateFound):** date, no value list, required. The date (yyyy-mm-dd) the researcher collected the error. This is not the air date of the audio recording, which is encoded in the audio file name. **File (recordFile):** text, no value list, required. File name of the audio recording of the podcast that the error is taken from. **Time stamp (recordTime):** time, no value list, required. Time stamp (h:mm:ss) in the audio file of the beginning of the longform entry that contains the error; this is not the location of the actual error, which is generally a little later. **Speaker (recordSpeaker):** text, no value list, required. A label identifying the talker that produced the error. Speaker initials are declared in the dataRecordings data table. **Sex (recordSex):** text, no value list, required. Sex {m, f} of the speaker. **Podcast (recordPodcast):** text, no value list, required. Label for the podcast series {ls, yg, ds}. **Confidence Intended (recordConfidenceIntended):** text, no value list, not required. “low” = the researcher has inferred the intended word or phrase but does not have 100% confidence in it; the intended is not obvious from context, but the researcher is sure the example contains an error. **Confidence Transcription (recordConfidenceTranscription):** text, no value list, not required. “low” = the analyst confirming the transcription is uncertain about some details. ===== **MAJOR CLASS FIELDS** The major class fields are required for every speech error and they constitute the core analysis of an error and follow standard speech error classification systems. **Master Type (classMasterType):** text, MasterType value list, required. Essentially a combination of the Level and Type values that gives a characterization of the error’s etiology. **Alt. MType (classAlternativeMasterType):** text, MasterType value list, required for ambiguous errors. When a speech error is ambiguous, the field Master Type gives the preferred analysis, and Alt. MType specifies an alternate analysis that is possible but not as likely as the preferred analysis. The Notes field typically explains the rationale for the preferred analysis. **Level (classLevel):** text, ErrorLevels value list, required. The basic linguistic unit of the intended term that is analyzed by the Type field. All phonological errors and phonetic errors are specified Sound, all modifications of at the level of the word are specified Word (e.g., Lexical Substitutions, Word Additions, Extreme Reductions), and Morpheme and Phrase levels involve operations on these units. **Type (classType):** text, ErrorTypes value list, required. Characterizes the operation on the linguistic unit of the intended term using a set of operations from standard types, with some additions, including gradient sound errors and extreme reductions. **Direction (classDirection):** text, ErrorDirections value list, required. Specifies the relationship of the error term to the source term(s) for contextual errors, and “NotAppl” for noncontextual errors; must be consistent with position of source words given in the longform of the error. Two additional fields, Direction 4Syll and Direction 7Syll, use different processing windows (4 and 7 syllables, respectively) to cross-classify directionality for a subset of the data (all tone errors, Mastertype=Tone Error). **Contextual? (classContextual):** text, YN value list, required. “Y” for contextual errors, “N” for non-contextual errors (which have Direction=NonAppl). **Right Lexeme? (wordLemmaError):** text, YNNotAppl value list, required. Classifies errors by determining if the right lexeme was used: “Y” if correct lexeme selected, as in sound errors (which are errors of phonological encoding, not lexical selection), “N” if not, as in lexical substitutions, and “NotAppl” in cases, like Word Additions, where the field does not naturally apply to the error. ===== **WORD FIELDS** These fields analyze the word-level properties of the intended and error, and the relationships between them. **Lexical Word? (wordLexicalError):** text, YNNotAppl value list, required. Specifies whether the error term is an actual word or not: “Y” = error is an actual word or phrase, “N” if error is not an actual word or phrase, “NotAppl” if lexicality is not relevant (e.g., word deletions). **Intended Part of Speech (wordPosIntended):** text, PartOfSpeech value list, required. Specifies the part of speech of the intended term from a fixed set, or “NotAppl” if there is no intended term (as in Word Additions), or “See Notes” if the intended cannot be straightforwardly categorized; word strings that constitute a phrase are assigned the part of speech of the head of the phrase. **Error Part of Speech (wordPosError):** text, PartOfSpeech value list, required. Specifies the part of speech of the error term from a fixed set, or “NotAppl” if there is no error term (as in Word Deletions), or “See Notes” if the error cannot be categorized; word strings that constitute a phrase are assigned the part of speech of the head of the phrase. **Intended Open/Closed (wordOpenIntended):** text, OpenClosed value list, required. Specifies if the intended is an open class item, closed class item; “NotAppl” for errors lacking an intended (e.g. Word Addition), and “See Notes” is used for word strings that are not phrases. **Error Open/Closed (wordOpenError):** text, OpenClosed value list, required. Specifies if the error is an open class item, closed class item; “NotAppl” for errors lacking an error term (e.g. Word Deletion), and “See Notes” is used for word strings that are not phrases. **E-I Semantic Relationship (wordErrorIntendedSemanticRelationship):** text, SemanticRelationship value list, required. Specifies the semantic relationship between the intended and error terms; “NotAppl” = one of the terms are missing, so there is no relationship; **E-I Morphological Relationship (wordErrorIntendedMorphologicalRelationship):** text, MorphologicalRelationship, required. Specifies the morphological relationship between the intended and error terms; NotAppl” = one of the terms are missing, so there is no relationship ===== **SOUND FIELDS** These fields analyze sound errors with respect to sound structure, including the segmental and tonal units, CV skeletal structure, syllable structure, word position, and phonotactic well-formedness. In the layout, the fields are organized into the same columns that specify aspects of the Intended (a.k.a., Supplanted Intended), the Error (=Intruder), and Source. **Supplanted Intended (soundSupplantedIntended):** text, no value list, required for sound errors. Specifies the intended phonological structure that is supplanted by the error; can be segments, tone, or “Ø” in the case of Addition; in sound exchanges, this is the first sound in the exchange. **Intruder (soundIntruder):** text, no value list, required for sound errors. Specifies the phonological structure that supplants the intended structure, or the intruder sounds; in sound exchanges, this is the second sound in the exchange. Deletions have “Ø” as intruders, indicating that the supplanted intended goes to null. The intruder sound is, by assumption, identical to the source sound, so the source sound does not require a separate field. **Supplanted Intended CV Structure (soundCVStructureIntended):** text, no value list, required for sound errors. Specifies the CV skeletal structure (abstract consonant-vowel positions in a structure frame), for the intended sounds. **Intruder CV Structure (soundCVStructureIntruder):** text, no value list, required for sound errors. Specifies the CV skeletal structure (abstract consonant-vowel positions in a structure frame), for the intruder sounds. **Supplanted Intended Syllabic Role (soundSupplantedIntendedSyllableRole):** text, SyllableRoles value list, required for sound errors. Specifies the syllabic role of the Supplanted Intended; “NotDiscernable” = the role is not clear, often because the word is clipped. **Intruder Syllabic Role (soundIntruderSyllabicRoleError):** text, SyllabicRoles value list, required for sound errors. Specifies the syllabic role of the Intruder in the error; “NotDiscernable” = the role is not clear, often because the word is clipped. **Source Syllabic Role (soundIntruderSyllabicRoleSource):** text, SyllableRoles, required for sound errors. Specifies the syllabic role of the Source sound; “NotDiscernable” = the role is not clear, often because the word is clipped. **Supplanted Intended Word Position (soundSupplantedIntendedWordPosition):** text, WordPositions value list, required sound errors. Specifies the position of the Supplanted Intended in a word, “Initial” if in beginning of word, etc.; “InitialAndFinal” is used when the string is the entire word; “NotDiscernable” = the position is not clear from the example. **Intruder Word Position (soundIntruderWordPositionError):** text, WordPosition value list, required for sound errors. Specifies the position of the Intruder in the error word, “Initial” if in beginning of word, etc.; “InitialAndFinal” is used when the string is the entire word; “NotDiscernable” = the position is not clear from the example. **Source Word Position (soundIntruderWordPositionSource):** text, WordPositions, required for sound errors. Specifies the position of the Source sound in the source word, “Initial” if in beginning of word, etc.; “InitialAndFinal” is used when the string is the entire word; “NotDiscernable” = the position is not clear from the example. **Supplanted Intended Whole Syllable (soundWholeSyllableIntended):** text, no value list, required for sound errors. Specifies the entire syllable that the Supplanted Intended occurs in; sometimes the syllable needs to be inferred from the context when the word is clipped (a kind of standardization). **Intruder Whole Syllable (soundWholeSyllableIntruder):** text, no value list, required for sound errors. Specifies the entire syllable that the Intruder occurs in; sometimes the syllable needs to be inferred from the context when the word is clipped (a kind of standardization). **Supplanted Intended Tone of Syllable (soundSupplantedIntendedSyllabletone):** text, no value list, required for sound errors. Specifies the tone that occurs on the syllable that contains the Supplanted Intended. **Intruder Tone of Syllable (soundSupplantedIntendedSyllabletone):** text, no value list, required for sound errors. Specifies the tone that occurs on the syllable that contains the Intruder. **Source Tone of Syllable (soundSupplantedIntendedSyllabletone):** text, no value list, required for sound errors. Specifies the tone that occurs on the syllable that contains the Source. **Error Action Is In (soundActionIsIn):** text, ActionInLocation value list, required for sound errors. Specifies the syllabic domain or domains of the error. **Identical Neighboring Segment? (soundIdenticalNeighboringSegment):** text, YN value list, required for sound errors. Indicates whether the Intruder sound appears in a phonetic context that is identical to the context for the Source sound; here “identical” means is there an identical segment before or after the Intruder/Source sound; see the phenomenon of the repeated phoneme effect in the literature; for deletion and addition, we look for identical segments in the input and output, respectively. **Phonologically Illegal? (soundPhonologicallyIllegal):** text, YN value list, required for sound errors. Indicates whether the error resulted in a phonologically illegal structure, that is, a phonological structure that is not allowed by Cantonese phonotactics; “Y” indicates the error is illegal and indicates “N” legal errors. **Triggers Resyllabification (soundTriggersSyllabificationYN):** text, YN value list, required for sound errors. Indicates whether, as a result of the error, some of the intended sounds are syllabified differently than they would have been without an error. ===== **SPECIAL CLASS FIELDS** Special class fields are used to describe the subtypes of specific kinds of data. **Obvious Malapropism (specialObviousMalapropism):** text, YNNotAppl value list, required for lexical substitutions. Indicates whether a lexical substitution is a malapropism, i.e., if there is a non-trivial phonological resemblance between the intended and error words. **Phonotactic Violation (specialPhonotacticViolation):** text, PhonotacticViolationType, required for illegal sound errors. Specifies the type of phonotactic violation in errors that violate phonotactic rules, e.g., illegal onset, illegal rime, etc. **S-blend T-map (specialSeqBlendToneMapping):** text, SeqBlendToneMapping list, required sequential blends. Specifies the way in which the tones of the intended are mapped to a smaller set of syllables in the error word of a sequential blend error; generally used to determine how tone can be permuted in these blends. **Gradient Type (specialGradientType):** text, GradientType value list, required for Phonetic Errors. Specifies the type of gradient phonetic error; see 5.1for explanation of the types. **Morpho Cats (specialMorphologicalCategory):** text, MorphologicalCategories, required morphological errors. Specifies the morphological category of the operative unit in a morphological error. **Check All Complex Processes (specialComplexProcessesAll):** text, ComplexProcesses value list, required for complex sets. Indicates all the processes involved in errors of type Complex Set of Processes.