File Information
File: 05-lr/acl_arc_1_sum/cleansed_text/xml_by_section/metho/80/c80-1092_metho.xml
Size: 21,028 bytes
Last Modified: 2025-10-06 14:11:18
<?xml version="1.0" standalone="yes"?> <Paper uid="C80-1092"> <Title>COMPUTATIONAL DIALECTOLOGY USING GLAPS --Automated Processing of Field Survey Data --</Title> <Section position="3" start_page="0" end_page="0" type="metho"> <SectionTitle> 2. Characteristics of GLAPS </SectionTitle> <Paragraph position="0"/> <Section position="1" start_page="0" end_page="0" type="sub_section"> <SectionTitle> 2.1 Easy Understandability </SectionTitle> <Paragraph position="0"> The GLAPS processor is a FORTRAN program of about 13,000 lines. It is a package program whose strongest point is that even people ignorant of computer programming can obtain output results using it.</Paragraph> <Paragraph position="1"> About thirty students of the Department of Linguistics, University of Tokyo, have used or are using GLAPS to produce crosstables from field survey data. (See, for example, Sapporo 1977 ~, 1978~.) Most of the students had never used a computer system before, but just a few hours of instruction were sufficient for them to understand how to use GLAPS and obtain their desired line-printer output.</Paragraph> </Section> <Section position="2" start_page="0" end_page="0" type="sub_section"> <SectionTitle> 2.2 Applicability to Various Data </SectionTitle> <Paragraph position="0"> GLAPS is applicable to various data, whether on fixed format cards, free format cards, or binary format disc files, and to any number of informants and variables or investigated items.</Paragraph> <Paragraph position="1"> The author and University of Tokyo colleagues have applied GLAPS to data in different formats from five field surveys (Shizukuishi 197~ I, 197~ 2, Tokunoshima 19763 ,</Paragraph> </Section> </Section> <Section position="4" start_page="0" end_page="0" type="metho"> <SectionTitle> 4 and Sapporo 1977 , 1978 ). Moreover, other </SectionTitle> <Paragraph position="0"> researchers have used GLAPS to process their own dialect data 6.</Paragraph> <Section position="1" start_page="0" end_page="0" type="sub_section"> <SectionTitle> 2.3 Compatibility with Various Computers </SectionTitle> <Paragraph position="0"> GLAPS is written in Japanese Industrial Standard (JIS) FORTRAN, level 7000, which is equivalent to Draft Recommendation FORTRAN of International Organization for Standardization at its maximum level (ISO Full FORTRAN) or ASA FORTRAN. It does not use assembly language and so is compatible with virtually all computer systems. In fact, GLAPS has been run on nine different computers without modification.</Paragraph> <Paragraph position="1"> 2.4 Flexibility with regard to Data Processing To run GLAPS, users simply prepare their dialect data and compose a short program written in 'GLAPS language'. (There are 75 different statements in this so-called language. Some of these appear from lines 2 to 75 in Fig. 3.) In this program, the user must specify all of the functions and operations to be performed. Most programs run only 20 to 30 lines, as we shall see below.</Paragraph> <Paragraph position="2"> GLAPS can perform a variety of functions needed for dialect data processing, such as the re-categorization of data, the pairing and combining of investigated word-forms, the deletion of unnecessary data, and the division of informants into subgroups by specified variables. Thus, GLAPS provides a versatile and flexible system for the user.</Paragraph> </Section> <Section position="2" start_page="0" end_page="0" type="sub_section"> <SectionTitle> 2.5 Processability of Multiple Answers </SectionTitle> <Paragraph position="0"> GLAPS resembles the SPSS (Statistical Package for the Social Sciences), originally developed at Stanford University. But GLAPS is capable of processing multiple answers often given to questions about word-form. The user simply specifies the number of answers to be accomodated in any given variable. GLAPS then automatically executes all statements related to the data and processes the specified number of answers.</Paragraph> </Section> </Section> <Section position="5" start_page="0" end_page="0" type="metho"> <SectionTitle> 3. An Example of the Application of GLAPS </SectionTitle> <Paragraph position="0"> As mentioned above, the author has applied GLAPS to several field studies. The following describes one of these.</Paragraph> <Section position="1" start_page="0" end_page="0" type="sub_section"> <SectionTitle> 3.1 Field Survey at Shizukuishi in 1974 </SectionTitle> <Paragraph position="0"> In 1974, a team from the Department of Linguistics, University of Tokyo conductedan intensive investigation to interview all the residents of the Nishiyama area of Shizukuishi township, lwate prefecture. The team interviewed 348 of about 500 residents above age 15, to examine distribution patterns of word-forms and the process of language change within a small area.</Paragraph> <Paragraph position="1"> : Fig. 2 Some Data Cards from Shizukuishi 1974 Fig. 1 is a map of Shizukuishi township, which is surrounded by mountains. The dot in the center of the map indicates the town of Shizukuishi, The rectangle at the top of the map indicates the Nishiyama area. The map shows the six bus routes of the township, equivalent to its main roads. In between the two roads at Nishiyama area runs a river from north to south. The investigated area covers the nine communities of Nishiyama, divided naturally into east and west by the river.</Paragraph> </Section> <Section position="2" start_page="0" end_page="0" type="sub_section"> <SectionTitle> 3.2 Data Stored in One Disc File </SectionTitle> <Paragraph position="0"> All the data gathered from interviews was coded and punched on 80-column data cards, and transfered onto a disc file. Fig. 2 shows some of these cards. Four data cards were prepared for each informant. The KZN cards contain information about an informant's attributes. The BIO, E35, and G31 cards include answers about language usage. Though three answer fields were allowed for each language usage question, most</Paragraph> </Section> </Section> <Section position="6" start_page="0" end_page="348" type="metho"> <SectionTitle> 73 ATLAS COWLICK 74 CONTROL COMMUNITY,NATIVE-OR-NOT 75 ATLAS COWLICK 76 =END </SectionTitle> <Paragraph position="0"> Fig. 3 User's Program for Analysis of 'Cowlick' informants gave only one or two answers to a question. Thus, on the BI0 card of informant I011, there are only twelve answers for the thirty possible answer fields.</Paragraph> <Section position="1" start_page="0" end_page="348" type="sub_section"> <SectionTitle> 3.3 User's Program and Output Results </SectionTitle> <Paragraph position="0"> Fig. 3 is a sample program, using GLAPS language, for analysis of the item 'cowlick' (the whirl of hair on the head). This figure is a fairly large program derived from many smaller programs which were used to analyze 'cowlick' trial and error.</Paragraph> <Paragraph position="1"> '=GLAPS' of line 1 is the top line of the program, and '=END' of line 76 indicates the end of the program. The lines starting with '*' are comment lines which the GLAPS processor ignores, --606-and so any useful notes or references can be entered here.</Paragraph> <Paragraph position="2"> The CASES statement of line 7 denotes the number of informants, here 348. The VARIABLES statement is from line 8 to ii. If a line starts with a space (such as lines 9, I0, and Ii), it means the line continues from the previous one. Names of different variables are listed in VARIABLES statements. Any words, letters, and symbols except ',', '=', '(', and ')' can be used for variable names. Unlike FORTRAN or COBOL, the length of variable names is not restricted. Lines 13 to 15 is another VARIABLES statement. But in this statement, a parenthesized three (3) follows every variable name. This means that these variables have three answer fields, that is, room for three different multiple answers to each question.</Paragraph> <Paragraph position="3"> The CASES and the VARIABLES (and the FO~AT of lines 17 and 18) are non-executable statements. The READ statement of line 16 orders GLAPS to read ALL variables (defined by the VARIABLES statements) from input device number '7' using FORMAT statement labeled'700. Input device numbers like this are associated with data files outside a program. The number '7' here refers to the data file of Fig. 2. The FORMAT statement of lines 17 and 18 specifies data format. This is similar to the standard FORMAT statement of FORTRAN.</Paragraph> <Paragraph position="4"> The TITLES statement of lines 19 and 20 gives the title of the output results, in this case, two lines in length. The title can be revised by means of a different TITLES statement if needed.</Paragraph> <Paragraph position="5"> 3.3.1 Map of Investigated Houses. The purpose of lines 21 to 35 is the production of a map showing the distribution pattern of the nine communities as well as informants' houses. The SIZE statement of line 24 indicates that the map size is 25 lines by 45 columns. The LOCATION statement of line 25 indicates which variables to use for location decisions. In this case, they are NORTH/SOUTH and EAST/WEST. The PRETITLES statement of line 26 and the POSTTITLES statement of line 27 indicate the character strings to be printed at the top and the bottom of the map, respectively. The DELETE statement of line 28 deletes informants with informantnumbers from 2 to 9, that is, it selects the first informant from each family.</Paragraph> <Paragraph position="6"> The NAMES statement of lines 29 to 32 identifies the meaning of numbers used in the coded data. For example, the code number l of line 29 indicates 'Tate', and so on. The SYMBOLS statement of line 33 assigns symbols (including numbers, as in this case) for the numbers of the data code, for the purpose of mapping the data. This allows for much greater flexibility of design. The ATLAS statement of line 34 is an )K*** iatenaive investigation at Hlahi~ama **** (5hizukulahl,</Paragraph> <Paragraph position="8"> X*** lntenaive investigation at Hlahi~ama **** (Shl=ukuiahi, lwete PrePS,) ii iiiiiiiiii!11 Iii11111111 II IIII III IIIIIIll CROSSTABULATIOH OF NATIVE-OR-HOT BY COWLICK iiiiiiiitlllllll IIIlllllllllll III Ililllll II COWLICK (the whirr oPS hair on the head) // rough c~aasltlca~ion // HATIVE-OR-HOT COUNT I 1 34 36 45 50 53 60 67 90 ROW PERCEHT I uzumakl meklz~m meklgurl maklboal makuta mmkurmbo maruhoal makuraz W taumuzl COLUMN PERCEHT I onzi ai umonzi .................. I ........ I ........ I ........ I ........ I ........ I ........ I ........ I ........ I ........ I I I 41 I 8 I ? I 1 1 12 1 15 I 64 I 4 1 20 I native I 23.84 I 4,BS I 4.07 I 0.58 I 6.98 I 8.72 I 37,21 I 2,33 I 11.63 I I 68,33 I 47,06 I 87,$0 I 33.33 I 86.67 I S7.69 I 56.14 I 80,00 I 83,33 I --I ........ I ........ I ........ I ........ I ........ I ........ I ........ I ........ I ........ I 2 1 ig I 9 I I I 2 I 6 1 II I 80 I I I 4 I non-natlve I 18.46 1 8.74 I 0.97 I i.g4-1 S,B3 I 10,68 I 48,54 1 0,9? I 3,88 1 I 31.67 I 52.94 I 16.50 I 66.67 I 33 33 I 42.31 I 43.86 I 20 00 I 16.67 I --I ........ I ........ I ........ I ........ I ........ I ........ I --~&quot; ..... I ........ I ........ I COLUMN 60 17 8 3 12 26 114 5 24 TOTAL E1.82 6.18 2,91 1 09 6,56 g.45 41.45 1,22 8.73 .................. I ........ I ........ I ........ I ........ I ........ I ........ I ........ I ........ I ........ I 11 2 I 6 i 5 I 0 I 0 I 8 I 4I 0 I 0 I 19 over 70 I 10.53 I 31.58 I 26.3@ I 0~ I 0 I 10.53 I 01.05 I 0, I 0. I 6.91 I 3.33 I 35.89 I 62,50 I 0. I 0. I 7.69 I 3,51 I 0. I 0. I --I ........ I ........ I ........ I ........ I ........ I ........ I ........ I ........ I ........ I @ I ? I 7 I 11 2 I 11 2 I 10 I 0 I 11 31 over 60 I 22.58 I @8.58 I 3.23 I 8.45 I 3 23 I 6 45 I 32.86 I 0. I 3.e3 I 11.27 I 11.67 I 41.18 I 12.50 I 66 67 I 5.56 I 7 69 I 8,7? I 0. I 4.17 I --I ........ I ........ I ........ I ........ I ........ i ........ I ........ I ........ I ........ I 3 I 9 I 3 I 0 I 0 I 5 I 4 I 15 I 0 I I I 3? over 50 I 24.32 I 8,11 I 0. I 0. I 13.51 I 10 81 I 40.54 I 0. I 2.?0 I 13,45 I 15.00 I 17.65 I 0, I 0, I 27.78 I 15.38 I 13.16 I 0. I 4,17 I --I ........ I ........ I ........ I ........ I ........ I ........ I ........ I ........ I ........ I 4 I ie I I I 2 I 1 I 10 I 11 I 31 I 3 I I I ?2 over 40 I 16,67 I 1,39 I 2,78 I 1.39 I 13.89 I 15,28 I 43,06 I 4.17 I 1,39 I 26,18 I 20.00 I 5.88 I 25.00 I 33.33 I 55,56 I 42,31 I 27,19 I 60.00 I 4.17 I --I ........ I ........ I ........ I ........ I ........ I ........ I ........ I ........ I ........ I 5 I 9 I 0 I 0 I 0 I 2 I ? I 23 I I I I I 43 over 30 I 20.93 I 0. I 0. I 0, I 4 65 I 16.28 I 83.49 I 2.33 I 2.33 I 15,64 I 15.00 I 0. I 0. I 0, I 11 11 I 26.92 I 20.16 I 20.00 I 4.17 I --I ........ I ........ I ........ I ........ I ........ I ........ I ........ I ........ I ........ I 6 I t0 I 0 I 0 I 0 I 0 I 0 I 211 11 15 I 47 over 20 I 21.28 I 0. I 0. I 0. I 0 I 0 I 44.68 I 2.13 1 31.911 17.09 I 16,67 I 0, I 0. I 0. I 0, I 0 I 18,42 I 20,00 I 62.60 I --I ........ I ........ I ........ I ........ I ........ I ........ I ........ I ........ I ........ I 7 I II I 0 I 0 I 0 I 0 I 0 I 10 I 0 I 5 I 26 over 10 I 42.31 I 0. I 0 I 0. I 0, I 0. I 38.46 I 0. I 19.23 I 9.45 I 18,33 I 0. I 0, I 0, I 0. I 0. I 8.77 I 0. I e0,83 I --I ........ I ........ I ........ I ........ I ........ I ........ I ........ I ........ I ........ I COLUMN 60 17 B 3 18 26 114 6 84 275 TOTAL 21,08 6.18 2,91 1,09 6.55 9,45 4145 1180 6,73 NO OF CASE5 . 348 Fig. 6 Z%~ intensive investigation at Nlshl~ama ~*% (Shlzukuiahl, lwate Pref.) mmmm= * wmmm mm mmwewmmetmmmm.mwmmmmmeemmm* CR055TABULATIOH OF COMMUHITY BY COWLICK * meem*mmem*mmememm mmememmi*e*mm mmmlmonm COWLICK (the whirl or halt on the head) // rough C/lassiflcation // COMMUNITV COUHT I 1 34 36 45 50 53 60 , 67 90 ROW PERCENT I uzumaki maklz~um maklgurl makiboai makure makurebo maruhoai makurazg t=umuzi ROW COLUMN PERCEHT I onzi sl umonzi TOTAL .................. I ........ I ........ I ........ I ........ I ........ I ........ I ........ I ........ I ........ I 11 4 I 0 I 0 I 0 I 8 I 11 g I 0 I @ I 18 Tare I 22.2@ I 0. I 0 I 0, I 11.11 I S.56 I 50.00 I 0, I 11.11 I 6,58 I 6,67 I 0. I 0. I 0. I 11.11 I 3.85 I 7.89 I 0. I 8.33 I --~ ........ I ........ I ........ I ........ I ........ I ........ I ........ I ........ I ........ I 2 I 3 I 11 0I 0 I 1I 3 I g I 0I 1I 18 Shinogamori I 16.67 I 5.56 I 0. I 0 I 5.56 I 16.67 I 50.00 I 0, I 5.56 I 6.58 I 8,00 I 5.88 I 0. I 0, I 5,56 I 11,54 I 7.89 I 0. I 4.17 1 --I ........ I ........ I ........ I ........ I ........ I ........ I ........ I ........ I ........ I 3 I ~0 I g I I I 2 I 2 I 10 I 14 I I I 6 1 65 Shinokarawe I 30.77 I 13.85 I 1.54 I 3.08 I 3.08 I 15.38 I 21,54 I 1.54 I 9.@3 I 23.64 i 33.33 I 52.94 I 12.50 I 66.67 I 11.11 I 38.46 I 1@,28 i 20.00 I 25.00 I --i ........ I ........ I ........ I ........ I ........ I ........ I ........ I ........ I ........ I 4 1 8 I 11 0 I 0 1 3 I 8 I 6 I 4 I 2 I 86 Hay@sake I 30.77 I 3.85 I 0, I 0. I 1154 I ?.69 I 23.08 1 15.38 1 7.69 I 9.48 I 13.33 I 5,88 I 0, I 0. I 16.67 I ?.69 I 5.26 I 80.00 I 8.33 I --I ........ I ........ I ........ I ........ I ........ I ........ I ........ I ........ I ...... --I 5 I 5 I 6 I 0 I 0 I 3 I 8 I 10 I 0 I 4 I 36 HAg@@hi-Haw@sake I 13.89 I 16.6? I 0 r I 0 I 8,33 I @2,62 I 2798 I 0. I 11.111 13.09 I 8.33 I 35,29 I 0 I 0. I 16 67 1 30,?7 I 8 77 I 0. I 16.67 I --I ........ I ........ I ........ I ........ I ........ I ........ I ........ I ........ I ........ 6 I 13 I 0 I I I 0 I 3 I 11 18 I 0 I 4 I 40 Kami-Shinozaki I 32,50 I 0. I 2 50 I 0, I 7.50 I @.50 I 45 00 I 0. I 10.00 I 14,5~ I 21,6? I 0, I I@.50 I 0, I 16.67 I 3.85 I 15 79 I 0. I 16,67 I --I ........ I ........ I ........ I ........ I ........ I ........ I ........ I ........ I ........ ? I 2 I 0 I 2 I 11 0 I 0 I 13 I 0 I @ I 20 5himo-Shinozaki I 10,00 I 0. I 10.00 I 5 00 I 0. I 0, I 65.00 I 0. I 10.00 I 7.87 I 3.33 1 0. I 25.00 1 33 33 I 0, I 0. I 11.40 1 0 I 8.33 I --I ........ I ........ I ........ I ........ I ........ I ........ I ........ I ........ I ........ I 81 1I 01 0I 0I @I 01 gI 0I 11 13 Hlgashl-Shlnozskl I 7.6g I 0. I 0. I 0 I 15,38 I 0, I 69.23 I 0. I 7.69 I 4.73 I 1.67 I 0. I 0. I 0. I 11.111 0, I 7,89 I 0, I 4.1? I --I ........ I ........ I ........ I ........ I ........ I ........ I ........ I ........ I ........ I 9 1 4 I 0 I 4 1 0 I 2 I 11 26 I 0 I @ I 39 HI@hi-Shine=ski I 10.26 I 0. I 10.26 I 0. I 5.13 I 8.86 I 66.67 I 0. I 5,13 I 14.18 I 6.6? I 0. I 50.00 I 0. I 11,111 3.85 I 02.811 0, I 8,33 I --I ........ I ........ I ........ I ........ I ........ I ........ I ........ I ........ ~ ........ I COLUMN 60 17 8 3 18 @6 114 S 24 2?5 Fig, 8 instruction for the output of a map.</Paragraph> <Paragraph position="9"> Fig. 4 is the output result of this ATLAS statement. This figure shows, the distribution of the nine communities, plotting the locations of all informants' houses.</Paragraph> <Paragraph position="10"> The NDELETE statement of line 35 of Fig. 3 cancels the effect of the DELETE statement of line 28, that is, the GLAPS processor begins to treat all the informants hereafter.</Paragraph> <Paragraph position="11"> the production of crosstables. This is the first of three steps in our analysis of 'cowlick'.</Paragraph> <Paragraph position="12"> The SUBTITLES statement of lines 39 and 40 gives a more detailed explanation of the meaning of a variable --in this case, COWLICK. The IGNORE statement of line 41 orders that those data codes for COWLICK indicated on this line be ignored.</Paragraph> <Paragraph position="13"> The RECODE statement of lines 42 to 44 is for re-categorization. In the original dialect data, informants' answers were coded separately from other variants. But by using this RECODE statement, a new code is substituted for the original and a variety of codes put together.</Paragraph> <Paragraph position="14"> The NAMES statement of lines 45 to 47 associates the new code numbers with specific word-forms.</Paragraph> <Paragraph position="15"> Lines 48 to 51 contain another pair of RECODE and NAMES statements. Originally an informant's age was coded using a five-digit system. If an informant were born in Feburuary 1941, for example, his code was 94102. Someone born in August of 1896 was coded 89608. Lines 48 to 51 classify all the varieties of informants' age into seven groups. The CROSSTABS statement of line 54 means</Paragraph> </Section> </Section> <Section position="7" start_page="348" end_page="348" type="metho"> <SectionTitle> CROSSTABS NATIVE-OR-NOT,COWLICK CROSSTABS AGE,COWLICK CROSSTABS COMMUNITY,COWLICK CROSSTABS PRIMARY-SCHOOL-NAME,COWLICK </SectionTitle> <Paragraph position="0"> and produces four crosstables.</Paragraph> <Paragraph position="1"> Figs. 5 to 8 are the output results of this CROSSTABS statement. According to Fig. 5, no great difference exists between native and non-native informants. Note that all these word-forms are used by native speakers as well as non-native speakers. This means that even word-forms borrowed from outside have a strong foundation in this area now.</Paragraph> <Paragraph position="2"> Fig. 6 shows differences by age-group.</Paragraph> <Paragraph position="3"> 'Uzumaki' and 'maruhosi' are used by all agegroups. But, primarily older groups use 'makizyumonzi', 'makiguri', younger groups use 'tsumuzi', and middle-age groups 'makure', 'makurebosi', and 'makurezyumonzi'.</Paragraph> <Paragraph position="4"> Fig. 7 shows differences by community. For example, 'makizyumonzi' and 'makurebosi' are more common in 'Shinokawara' and 'Higashi-Hayasaka', and so on.</Paragraph> <Paragraph position="5"> In Fig. 8, 'Nagayama' Primary School can be regarded as equivalent to the east side of the river and 'Nishine' to the west side.</Paragraph> <Paragraph position="6"> These figures reveal that each word-form has its own distribution pattern.</Paragraph> <Paragraph position="7"> to examine the combined influence of age and geography on the word cowlick: by linguistic maps and by glottograms. Lines 55 to 61 of Fig. 3 are instructions for producing linguistic maps classified by age.</Paragraph> <Paragraph position="8"> The CONTROL statement of line 59 instructs GLAPS to divide informants into subgroups by AGE and to print out maps for every age group.</Paragraph> <Paragraph position="9"> Since AGE was receded into seven categories on lines 48 to 51, seven maps of COWLICK--Figs. 9 to IS --are produced by the single ATLAS statement of line 60.</Paragraph> <Paragraph position="10"> Fig. 9 is for persons over 70, Fig. i0 for persons over 60, and so on. Fig. 9 shows a clear contrast between east and west. The eastern part uses 'makizyumonzi' whereas the western part 'makiguri'. These seven maps show a great difference between east and west. This suggests that glottograms of both sides of the</Paragraph> <Paragraph position="12"> (the whirl of hair on the head)</Paragraph> <Paragraph position="14"/> <Paragraph position="16"/> </Section> class="xml-element"></Paper>