User Guide

What is Colligator? ······································································································· 2 System Requirements ·································································································· 3 How to use Colligator? ································································································ 4 Results Display ·············································································································· 9 Limitations ····················································································································· 10 Copyright ························································································································ 11

Colligator User Manual

What is Colligator? The term colligation as used in this manual conforms to Firth (1957), Hoey (2004) and Stubbs (2001). Put simply, colligation is the grammatical company a word (e.g., can) or a word category (e.g., VM) keeps. Both collocation and colligation are types of phraseologism (Gries 2007). The difference lies in that collocation is concerned with the lexical company a word keeps, while colligation is concerned with the grammatical company a word or a word category keeps.

Colligator is a software program for corpus-based language research. It is particularly useful for analyzing colligations in POS-tagged texts.

The motivation for our software program is that the retrieval of colligational patterns in texts is much more difficult than that of collocational patterns. With Colligator and POS-tagged texts, this may be a much easier job.

2

System Requirements Colligator runs under Microsoft Windows (98, 2000, XP, Vista). In order for Colligator to run properly, you need to have

1) Microsoft .Net Framework 2.0 or higher installed on your computer (downloadable from http://www.microsoft.com/downloads/details.aspx?FamilyID=0856EACB-4362-4B0D-8EDD-AAB15C5E04F5&displaylang=en); 2) Part-of-Speech (POS) tagged texts to be analyzed.

The following is the text format (POS-tagged with CLAWS71) acceptable to Colligator:

Colligator_NN1 is_VBZ a_AT1 software_NN1 program_NN1 for_II corpus-based_JJ language_NN1 research_NN1 ._. It_PPH1 is_VBZ particularly_RR useful_JJ for_II analyzing_VVG colligations_NN2 in_II POS-tagged_JJ texts_NN2 ._. In such POS-tagged texts, words and POS tags are connected with underlines (e.g., Colligator_NN1).

1

All POS tags used in this manual are from CLAWS7.

3

How to use Colligator? After Microsoft .Net Framework 2.0 or higher is installed in your operating system, simply double-click Colligator’s program icon to run the application.

Colligator’s Interface

When the program is activated, you will see the above graphical interface. Follow these steps to retrieve colligational patterns from POS-tagged texts: 1) Select files Click the

button to browse your POS-tagged texts. Choose the texts

you want to analyze (you can use the Ctrl key or the Shift key to choose multiple texts). Click the OK button to confirm your selection;

4

Choose Texts

2) Define your search item

- If you are studying the colligations of a word token or a lexical phrase, you should then key in that word token or lexical phrase as the search item (e.g., can, you can). Choose Word as your search mode.

The Word search mode

- If you are studying the colligations of a POS tag or a POS tag sequence, you simply key in that POS tag or POS tag sequence (e.g., VM, PPY VM). Choose TAG as your search mode.

The TAG search mode

5

- If you are studying the colligations of a word_TAG combination or combinations, you simply key in that combination or combinations (e.g., can_VM, you_PPY can_VM). Choose Word_TAG as your search mode.

The Word_TAG search mode - Colligator supports the powerful regular expressions. Therefore, if you want to study the colligations of a category rather than a specific word or tag or word_tag combination, you can use regular expressions to make your search item more inclusive. For example, in the Word search mode, the regular expression \\w+ments? matches all words suffixed with -ment and -ments. In the TAG mode, the regular expression V\\w+ matches all verbs. In the Word_TAG mode, (make|made|makes|making)_VV\\w+ matches all verb forms of MAKE, but not making_NN1 nor makings_NN2. You can even use \\w+_\\w+ to search for all words in the texts you have chosen. (Note: \\w may match any alphanumeric sequence, but not contractions with an apostrophe ('). For example, to search for not and its contracted form n't, you probably need to use the expression [\\w']+_XX. Besides, to include hyphenated words in your search, you need to use the expression [\\w-]+). A complex regular expression to search for passive structures may look like [\\w']+_VB\\w+(?:[\\w']+_(X|R)\\w+\\s)*[\\w-]+_V\\wN, which matches not only standard passives like been reported (been_VBN reported_VVN), but also extended passives like isn't always believed (is_VBZ n't_XX always_RR believed_VVN).

Using regular expressions

3) Define your search span

6

If your search span is set to 0 (Left) and 0 (Right), you are searching for the words or tags or the combination of both without looking at the context. A rule-of-the-thumb is to set the Left context to 0 first (to look into the colligational behavior of the search term in its right context) and then to set the Right context to 0 (to look into the colligational behavior of the search term in its left context).

Displaying the colligational patterns of the search term in its Left context

7

Displaying the colligational patterns of the search term in its Right context

4) Run the program

After all your settings are done, click the Run button to run the program.

8

Results Display

Colligator has three windows to display its results. As shown in the above picture, the upper left window displays the colligational patterns of the search term in its context. If the search term is a TAG or a Word_TAG, the Details window displays the results in more detail, while the window on the right displays the matches in the most detailed manner (the actual occurrences in the texts), so that the viewer can see which lexical strings match the pattern(s) in the upper left window. If your search term is a word or a lexical phrase, the window on the right will not show anything, as the most detailed results will be shown in the windows on the left.

If you would like to copy the results, simply right-click any of the lines in the appropriate display window, choose “Copy Results”, and paste the clipboard into any text processor.

Copy results to clipboard

9

Limitations

If you are working on really large text files, the retrieval process may take quite some time. It may even be a good idea to make some tea while waiting. The time taken depends on how large your texts are and how much memory your computer has available.

The real limitation for this version of Colligator is that the number of files to be searched in should not exceed 2,184. If it is necessary to search in more files, it is suggested that some or all of the texts be merged so that the total number of texts does not exceed 2,184.

10

Copyright

Colligator is programmed by Dr XIONG Wenxin, and designed by Dr LIANG Maocheng and Dr XU Jiajin, all working at the National Research Center for Foreign Language Education, Beijing Foreign Studies University. Bug reports are highly appreciated.

Colligator程序由熊文新博士编写，梁茂成博士、许家金博士设计。以上人员皆为北京外国语大学中国外语教育研究中心研究人员。如果您在使用过程中发现本软件存在bugs，请报告。

版本更新信息请查外语教育研究中心外语教研网和www.corpus4u.org。 Citation:

熊文新、梁茂成，类联接分析工具Colligator在基于语料库的语言研究中的应用，《中国外语教育》2008年第3期

11