Policy about Use of Dictionaries and Automatic (Machine, On-Line) Translators

last modified: 9/22/11

Here are, first, a statement of policies about use of language resources, and then some explanations and advice. The policies are "default" rules; it may be that a specific activity will include specifications that adjust the policies. You should also note that, just because a policy that is stated here does permit use of a resource, that does not mean that we recommend it. English-to-other-language dictionaries often cause more problems than they solve, at least until people have learned enough of the target language that they can sense whether the word they have selected from several possibilities is indeed likely to be appropriate.

activity dictionary computer translator INTO English computer* translator translator FROM English
writing test, whether inside or outside the classroom NO NO NO
individual project YES YES NO
reading test, whether inside or outside the classroom NO NO NO
listening test, whether inside or outside the classroom NO NO NO

About human translators and other help from human beings (other than the instructor): This is a gray area. Light editing? Probably OK, and probably effective for learning purposes. Telling, in English, an advanced or native speaker (family member, etc.) what you want to have written in the target language, and then reproducing closely and at length what that person says or writes? NO.

The policy established here is based on 3 principles:

1) It is dishonest and academically unacceptable to claim as your own work what has been done by someone (or something!) else. This applies especially to use of on-line translators.

2) The most important purpose of our course is to learn a language. If you use a resource that allows you to short-cut the learning, you are defeating the purpose of the course, particularly as it applies to your own learning.

3) Resources like dictionaries and on-line translators might be appropriate for some activities and inappropriate (or dishonest) for others.

Let's allow for some relativism here:

Regarding principle 1: If you write something in English all by yourself, then you didn't steal that - the English text. Outside a language class, under most circumstances no one would object if you used a dictionary and on-line translator to convert what you wrote to another language. And we think you should be taught how to use those resources, which we will do. But it is dishonest to turn in something translated by someone or something else, as though it is your own composition in the language you are learning.

Regarding principle 2: In the real world, people use dictionaries and on-line translators. An employer would be rather upset if an employed failed to use such resources. So (see above) we will help you learn to use them. But you simply cannot subsitute using those resources for learning how to understand and write the language you are learning. Your self-interest plays a role here: if you use dictionaries and, especially, on-line translators when you can to do your writing for you (or help you understand the language), you will likely pay dearly for that when you take a test where those resources are not allowed.

Regarding principle 3: Since we are not prohibiting dictionaries and on-line translators for each and every activity, the rule for using or not using them will be stated, either here as basic policies, or with the specification for the individual activities. Along the way we will also give advice about when and how to use such resources.

So how will we enforce these policies, and how can we detect violations of them?

1) Overt cheating (example: using dictionary during in-classroom exam) will be dealt with in the usual manner as a violation of the honor code. This has never been a serious problem.

2) The chief problem is likely to be use of on-line translators (or human resources!) to substitute for a student's own writing on a project, or on a test that is done outside the classroom. When we think that may have occurred, we will proceed cautiously but resolutely. If we think there is good reason to believe that prohibited resources have been used, we will refuse to accept the work and will insist that it be redone legitimately. Unless the matter is extreme (repeated violations of policy), we will not pursue the issue officially. If a student errs here in good faith, we will just advise, teach and assist.

How do we approach this as caring language teachers?

Before suspecting that inappropriate resources have been used, we have to consider whether the student has actually written the piece but has previously acquired language skills and should be placed in a higher-level course. We try to deal with that early. The real problem, though, is when we detect that the writing simply cannot have been produced by the student without some significant external assistance. Here we can usually tell what has happened. Here are the various distinguishing characteristics:

a) Texts produced by machine translators will show a high level of grammatical accuracy, even when what they produce is otherwise nonsensical. There will be no spelling errors. Vocabulary may be incorrect where an English word has several possible equivalents of which only one is actually right in the given instance. (Example: "drive" can mean one thing in psychology, another in computers, and still another in cars.) The "human touch" will be absent. If the starting (=English) text is casual (or ungrammatical), the translated text will likely be gibberish.

b) Texts produced by native-speaker translators will show a high level of grammatical accuracy, and also of vocabulary choice. But sometimes they will include native-speaker errors in spelling, or of grammar (substandard structures). They may include "flavor words", idioms, or slang that machine translators and non-native speakers do not handle. Accuracy in details of grammar will likely exceed the competence of non-native students, even those who have majored in the language, or even of many graduate students.

c) Texts produced by non-native-speaker translators, unless they are extremely fluent, will show a relatively high level of attention to "textbook" grammar, even in structures that are far beyond the level of students in the "writer's" course. If the topic is casual, the vocabulary may not be as "casual" as it should be, and "flavor words" may not be used, or not used naturally. Vocabulary will likely be wildly above what has been introduced in the "writer's" course. Native-speaker errors will not occur. Often the tone or "register" will strongly indicate that a text has been translated, not simply re-expressed or created within the target language.

In short: we can tell when it's your writing, and when it's not.