Linguistics 483 Data Files

A. C. Brett acbrett@uvic.ca Department of Linguistics University of Victoria Clearihue C139

Data Files

• dictsamp: a quasi-random systematic sample of words from the Nelson Highroads Dictionary, New Canadian Edition.
• clausamp: contains information on the individual sentences comprising fragments of English-language text from several different sources.
• babytalk: a sample of lengths of 100 utterances by an adult to a three-year-old infant from Woods, Fletcher, and Hughes (1986:14).
• saffran: a sample with approximately the same parameter estimates as those reported by Saffran et al. (1996) in their paper "Statistical learning by 8-month-old infants."
• teenty: data collected, and provided to us, by Robert J. Smith, a Linguistics 483 student who was studying the acoustic differences between "-teen" and "-ty" for a Linguistics 380 project.
• caneng: data from the pilot phase of a project to study Canadian English usage and social attitudes.
• npfreq: a sample with approximately the same parameter estimates as those reported by Alario et al. (2002) in their paper "Frequency effects in Noun Phrase production: Implications for models of lexical access."
• teacher: data from Tables 10.1 and 11.5 on pages 155 and 186, respectively, of the text, Woods, Fletcher, and Hughes (1986). See page 154 for a description of the data and their source.
• gramvoc: data adapted from Figure. 1 reproduced in Bates, Devescovi, and Wulfeck (2001). See Bates et al. for this figure and reference to the orginial data.

File Formats

• dictsamp.txt: a quasi-random systematic sample of words from an elementary school dictionary consisting of observations on five variables as follows:

      Variable     Position    Description and Coding
      ________     ________    ___________________________________
  
      WORD         1 - 19      Dictionary head word, coded as
                               alphanumeric
  
      LANGUAGE     20          Etymological code identifying the
                               language from which the word entered
                               English, coded alphanumeric as:
                               
                               'L'  Latin
                               'G'  Greek
                               'F'  French
                               'A'  Anglo-Saxon
                               'H'  Hindi
                               'P'  Persian
                               'S'  Scandinavian
                               'I'  Italian
                               'C'  Spanish
                               'E'  English
                               'D'  Dutch
  
      LETTERS      21 - 24     Number of letters in the word, coded
                               as integer
  
      SOUNDS       25 - 28     Number of phonetic symbols in the
                               word, integer
  
      SYLLBLES     29 - 32     Number of Syllable in the word,
                               integer
  
  

• clausamp.txt: consists of observations on four variables. The variable names and the positions in the data records of the observations on these variables are given in the following table. The table also contains a description of the variables and of the nature and coding of the observations recorded on them.

      Variable     Position    Description and Coding
      ________     ________    ___________________________________
  
      TYPE          3 - 4      Source document type, coded as:
  
                               '01', if the sentence be from
                                     an informal narrative document
                                     (newspaper or magazine), or
                               '02', if the sentence be from a formal
                                     narrative source (technical
                                     article or monograph).
  
      WORDS         5 - 8      The total number of words in the
                                     sentence, coded as integer.
  
      LETTERS       9 - 12     the total number of letters in the
                                     words comprising the sentence,
                                     coded as integer.
  
      STRUCTUR     15 - 16     Sentence structure, coded as:
  
                               '00', if the sentence be of simple
                                     structure (one independent
                                     clause), or
                               '01', if the sentence be compound (one
                                     or more independent clauses, but
                                     no dependent clause(s)), or
                               '10', if the sentence be complex (one
                                     independent clause and one or
                                     more dependent clause(s)), or
                               '11', if the sentence be compound/
                                     complex (two or more independent
                                     clauses and at least one
                                     dependent clause).
  

• babytalk.txt: consists of observations on two variables, an utterance seqence number and the utterance length in morphemes, as follows:

      Variable     Position    Description and Coding
      ________     ________    ___________________________________
  
      SEQ.NUMB     1 - 3       Utterance Sequence Number, integer
  
      LENGTH       4 - 6       Utterance Length in Morphemes,
                               integer
  

• saffran.txt: Observations on two variables, both of which correspond to head turn durations, made on 8-month-olds by Saffran et al. (1996) and which were employed to support the claim that infants can segment words from fluent speech strictly on the basis of statistical information.

  Observations on the first variable, named FAMILIAR, represent head turn durations elicited by pseudowords that were presented during the two-minute of the habituation phase of the experiment. Values on the NOVEL variable correspond to head turn durations resulting from pseudowords made up of the same syllables, but presented in different orders from the pseudowords in the habituation corpus.

  Note that the data used here are not the Saffran et al. data. The data used in this example consist of a pseudorandom sample generated with approximately the same parameter estimates (statistics) as those reported by Saffran et al.

                       Position
       Variable        in Data          Attribute Measured
       Name            Records          or Identified
       ________        ________         ________________________
  
       FAMILIAR          1 - 8          Familiar items head turn
                                        times
  
       NOVEL            9 - 16          Novel items head turn
                                        times
  

• teenty.txt: Measurements made on six subjects (identified by single letters in the data file), each of whom read a collection of seven-digit numbers. Each of these numbers consisted of a single digit, followed by three two-digit pairs. Subjects were asked to read, and utter, the first digit (of each seven-digit number), followed by the subsequent three two-digit pairs, as a unit. Thus, for example, "60" would be read as "sixty," and "16" was read as "sixteen." Each target number consisted of the root, followed by a "-ty" or "-teen" affix, and were located in the seven-digit context in either the last, or second to last position.

  Fundamental frequencies (in hertz), and the start and end times (in hundredths of milliseconds), of subjects were measured as they uttered the target two-digit numbers. Initial, midpoint, and final fundamental frequencies of the root number and the affix were recorded to yield six frequency observations (three for the root morpheme, and three for the affix), followed by four times (start and end times for each of the root and the affix, measured in hundredths of milliseconds), for each target number. Positions of these observations in the data records, together with the speaker identifier, the context (seven-digit number), the position of the target number in the context, and the target number itself, are as follows:

                       Position
       Variable        in Data          Attribute Measured
       Name            Records          or Identified
       _______         ________         ______________________
  
       SPEAKER            1(A)          Speaker ID letter
       CONTEXT          3-9(A)          Complete number spoken
       POSITION             11          Target number position
       NUMBER            13-14          Target number
       RF1               16-18          Root F0 at start
       RF2               20-22          Root F0 at midpoint
       RF3               24-26          Root F0 at end
       AF1               28-30          Affix F0 at start
       AF2               32-34          Affix F0 at midpoint
       AF3               36-38          Affix F0 at end
       RT1               40-44          Root V start time
       RT2               46-50          Root V end time
       AT1               52-56          Affix V start time
       AT2               58-62          Affix V end time
  
  

• caneng.txt: A data file derived from the results of a pilot study for a project to survey Canadian English usage and social attitudes.

  Participants in the pilot project consisted of 64 teachers taking introductory linguistics courses during Summer Session.

  The survey instrument included 405 items. For purposes of the analyses being undertaken, responses to related items were combined to produce measurements on 31 scales. Since different numbers of items were included in each scale, the scales were transformed so that the values on each of them ranged from 0 to 99.

  The data file stored here consists of the measurements on these 31 transformed scales together with responses on the biographical variables.

      Variable   Position   Description
      ________   ________   ________________________________________
  
        AGE       1-3       Informant Age in Months
        SEX       4         Informant Sex
        GENER     5         Generation in Canada
        STATUS    6         Father's Occupation Socioeconomic Status
        S1A       7-8       Anomie
        S1B       9-10      Ethnocentrism
        S1C       11-12     Preference for Canada over USA
        S1D       13-14     Preference for US over UK
        S1E       15-16     Preference for Canada over UK
        S1F       17-18     Attitude toward USA Relative to the UK
        S1G       19-20     Authoritarianism
        S1H       21-22     Sensitivity for Others
        S1I       23-24     Canadian Nationalism & National Identity
        S2A       25-26     Attitude toward Americans
        S2B       27-28     Attitude toward Scots
        S2C       29-30     Attitude toward English
        S2D       31-32     Attitude toward Canadians
        S2E       33-34     Subject Self-Rating: Others See Her/Him
        S2F       35-36     Subject Self-Rating: Self Perception
        S3A       37-38     Morphosyntactic: What the Subject Hears
        S3B       39-40     Morphosyntactic: Use with Superiors
        S3C       41-42     Morphosyntactic: Subject Uses with Peers
        S4A       43-44     Phonological: Vowel before "R"
        S4B       45-46     Phonological: Syllabicity
        S4C       47-48     Phonological: Low Back Vowel
        S4D       49-50     Phonological: Diphthong Raising
        S4E       51-52     Phonological: "Y"-Glide
        S4F       53-54     Phonological: "WH"-Feature
        S4H       55-56     Phonological: Vowel Deletion
        S4I       57-58     Phonological: "T"-Flapping
        S4J       59-60     Morphophonemic: US/UK Dichotomy
        S4K       61-62     Morphophonemic: Standard/Nonstandard Dialect
        S4L       63-64     Lexical: UK/US Dichotomy
        S4M       65-66     Lexical: Standard/Nonstandard Dialect
        S4N       67-68     Lexical: Canadian Local Dialect
  

• npfreq.txt: Data synthesized by a pseudo-random number generation process to obtain a sample with approximately the same parameter estimates as the observations collected by Alario, Costa, and Caramazza (in press) in an experiment to assess the effect of noun and adjective frequency on noun phrase production latencies.

  Note that the data used here are not the Alario et al. data.

      Variable   Position   Description
      ________   ________   ______________________________
  
      NOUN        1 - 8     Noun (alphanumeric)
      NFREQ      10 - 12    Noun Frequency
      ADJ        15 - 22    Adjective (alphanumeric)
      AFREQ      24 - 26    Adjective Frequency
      I_NFREQ    29 - 30    Noun Frequency Indicator
      I_AFREQ    33 - 34    Adjective Frequency Indicator
      TIME       37 - 41    NP Production Latency
  

• teacher.txt: Data consisting of assessments by three groups of judges of the severity of errors in 32 sentences produced by Greek-speaking learners of English.

  The judges consisted of ten English teachers who were native speakers of English, ten native speakers of English who were not teachers, and ten English teachers whose native language was Greek. Each of the 32 sentences contained one error, and the judges evaluated the severity of the error on a 0 - 5 scale. For each sentence, the scores for each group of judges consisted of the sum of the severity assessment scores for the group.

  The data are comprised of the sentence number and the total severity scores for the sentence determined by each of the three groups of judges.

      Variable   Position   Description
      ________   ________   ______________________________
  
      SENTENCE    1 - 2     Sentence number
      E_TEACHR    3 - 6     English-speaking teacher
      G_TEACHR    7 - 10    Greek-speaking English teacher
      NONTEACH   11 - 14    English-speaking non-teacher
  

• gramvoc.txt: Data consisting of the Grammatical Complexity Scores and Vocabulary Levels for English- and Italian-speaking children.

  The Vocabulary Level corresponds to the size of the vocabulary of the children, and the Grammatical Complexity Scores for the English-speaking and Italian-speaking children measures the grammatical complexity of their utterances as assessed by their caregivers.

      Variable   Position   Description
      ________   ________   _______________________________
  
      VOCAB       1 - 4     Vocabulary Level
      SYN_ENGL    5 - 10    Grammatical Complexity (English)
      SYN_ITAL   11 - 16    Grammatical Complexity (Italian)