One of the things that I have found most difficult and puzzling about teaching is how to grade translations. It is hard to balance objectivity and subjectivity. Furthermore, at what level do you expect students to try to mimic the author’s word order? How do you account for the translator’s writing style? To what degree can we even say that a translation is “correct” when we have to move thoughts from 2000 years ago into today’s vocabulary and mindsets? For the most part, the answer to the last question must be that a translation actually needs to reflect an understanding of Latin grammar–does the translation maintain the passive voice? do all nouns retain their singular/plural quality from the Latin? do the English adjectives modify the nouns that they modify in Latin?
When I first started teaching Latin, I attempted to be incredibly objective and take off half a point for each mistake in the translation. As the term progressed, I quickly realized a flaw in this system. If the student thought the subject was singular, then the verb will also be singular. Is this one or two mistakes? If the student switched the subject and direction object in their translations, is this one or two mistakes? How do you adjust the grade to reflect mistakes that you clearly understand resulted in other mistakes? I began to only take off half a point for some of these translations, but it still seemed like a flawed grading system. For example, is it more important that the student translated the verb passively in a quiz about the passive voice than that the subject was singular instead of plural? How many points can you take off for each word that might have several problems?
Next, I attempted to deal with these issues by considering what I considered to be “minor” and “major” mistakes. A minor mistake could be confusing singular words for plural or using a meaning for a vocabulary word that was accurate but not the best meaning in this sentence. A major mistake was confusing the voice or tense of the verb or confusing the case of a noun. With these in mind, I developed this basic idea for how to grade students for a sentence worth 10 points:
- 10 points for perfect, grammatically correct sentence
- 9 points for 1-3 minor errors (singular/plural, alternate meaning of word or idiom)
- 8 points for 1-5 minor errors or 1-2 major errors (case usage, tense, voice, mood, meaning of word)
- 7 points for 1-7 minor errors and/or 1-3 major errors
- 6 points for 1-9 minor errors and/or 1-5 major errors
- 5 points for effort – whole sentence translated but incorrectly
- 3 points for partially complete translation
As I used this scheme, I realized a problem: minor mistakes were giving a student a 90% (or the borderline between an A and a B on my grading scale). This seemed harsh so I began mainly giving students 9.5 or 8.5/10 rather than the even 9/10. Another problem arose: I was writing short sentences for students to translate so that I could ensure that I was testing that chapter’s concept. Short sentences eliminated the utility of this seemingly objective grading scheme. If a student completely translated Carthago delenda est but with several major errors, did that count as 7, 6, or 5 points?
Therefore, currently, I grade students based on a more subjective scale. I determine whether their translation merits an A, B, C, D or F. The grading scheme that I use, then, is:
- 10 points for perfect, grammatically correct sentence
- A = 9.5 points for minor errors that show a clear understanding of the sentence, but with a slightly off translation of one word (i.e. vocabulary meaning may be wrong, a singular for plural, and possibly even a tense being off).
- B = 8.5 points for a translation that shows the student understands the basic sense of the sentence but is struggling to put the translation together. There are more “major” errors but the student nearly understands the grammar concept tested in this sentence. If a student writes a blank instead of a translation of a word, this is usually the highest they can earn.
- C = 7.5 points for a translation that shows several major problems, has some potential to understand the grammar concept tested in this sentence but is not quite there yet, and/or has some problems with the voice or mood of the verb that are preventing the sentence from being properly translated.
- D = 6.5 points for a sentence that has some vocabulary or cases correct but shows major confusion in terms of piecing the puzzle of the sentence together.
- F = 5.5 (though usually actually 5 or 6) points for a sentence that shows very little understanding of the Latin. Usually, only one or two words is even in the correct case or has a proper translation of the word’s meaning.
- Incomplete translations earn a varied number of points based on how complete the sentence is and how accurate the complete parts are.
This scheme is not perfect. It is more subjective and can make it hard for students to understand where their grades come from, but I think it does the best so far of allowing a teacher to give a fair grade that does not penalize a student for too many small mistakes or reward them for only having one or two major mistakes. After all, how can you be perfectly objective grading a language when language and grammar are more subjective ideas?
I will point out two potential flaws with it: it works better with a scale of 10 points/sentence. When I have asked students to translate a passage on a test, I have tried not to make the passages worth too many points in relation to other parts of a test. In doing so, I have used the same system but made each sentence or clause worth 5 points. While it is simple to divide the above points by two, I have generally been reluctant to award students quarter points so that calculating the final grade is easier (especially after test corrections). Secondly, it makes test corrections a little more difficult for students. They are not always sure why they have not earned full credit for their translations. However, making students work a little harder to find their mistakes is fully within the spirit of offering test corrections, so I am little less concerned with this potential flaw.
I would be more than happy to hear other people’s strategies for grading translations and how to make translation grades as fair as possible.