Thanks to everyone for the comments on my plagiarism post; none of them were shocking, but it's always nice to get a broader perspective. My own response was captured rather completely by Rebecca and tenaciousmcd. Sometimes a sentence can slip through accidentally, but if one takes an entire paragraph verbatim from a source that does not appear anywhere in the paper, then that's the line for me. So 1, clearly.
As for penalties, I think I'll be very harsh on my very good friends out there that suggest anything less than C. I respect your opinions and arguments, but I'll say to all of you what I just told my colleague here: you're working in an old paradigm. In the paradigm that characterized most of our college experiences, plagiarism was this unbelievable act of obvious cheating that involved finding a scholarly source in the library and copying from it. It was, I think, something that happened fairly rarely, and it was certainly something that required a great deal of work to accomplish. Overall, it was abnormal. In the new paradigm, cutting and pasting from websites is normal for a very high percentage of students. It's just what they do: they google for their topic, surf to a bunch of websites, cut and paste, and then rearrange and add a few sentences of their own. (Of the cases I've just found in the last two weeks, the average number of different cites from which the student had stolen was 5.) It is not an accident.
The NY Times ran a survey a few years back and found more than 75% of US College graduates admitted to some form of cheating when they were in College. They are cheating and they know it's cheating. It's just that cheating has become normal. There was also a study done by a rat choice political scientist (I know) that showed (I think convincingly) that if the penalty is anything less than C it is rational to cheat. For all the dozens of cases I catch, there are dozens and dozens more that slide through.
I've had students cry in my office and swear it was an accident, but I'm now completely convinced that it never is. To Mark: students don't copy from one another (it's too much work) and if they do, I probably won't catch it so it's sort of a moot point. To Dan: you have a wonderful heart to see this issue the way you do (mine, as all can tell has become hardened beyond belief through the wrenching interactions with the long, long list of students that cheated), but the 'meant to go back' is always a cover. When a student puts together 50% of their paper verbatim from someone else's words, there would be nothing to go back to. When you take an entire paragraph from another source, put it in your paper and don't put quotation marks around it, you've already decided to cheat (even if you haven't admitted it to yourself, and the 'meant to go back' is the rationalization you give yourself).
And Rebecca makes a good point here, one that I hope contributes to my case for a new paradigm: they are NOT stealing from good sources, they are stealing from crap summary pages on the internet that sound, often, a lot like something they'd write anyway. My hunch is that in the old paradigm students would steal to get an A, steal because they thought there work was B work but they wanted it to be perfect so othey steal from a journal article. Now, we have B and C students stealing in order to get B's and C's; it's not about getting better grades so much as it is just plain easier to cut and paste than it is to write.
And finally, students have no fear at all of the repurcussions. Their logic looks like this...chances are I won't get caught, if I do I'll probably get to rewrite, worst case scenario is a bad grade in the course, but then I can just retake the course and substitute the new grade. Tmcd is right: in practice, it's almost impossible to go beyond C, because university institutions are set up to make it almost impossible to do much to the student. Combine this with the fact that a large number of professors let dozens of plagiarism papers slide off their desk each semester, and I see it as essential that faculty start imposing the harshest penalties possible.
And it can't just be done on a case-by-case basis, because it really is too much work. I've spent about 60 hours marking essays and exams over the past two weeks. I've spent another 10 hours loooking for, and when found, documenting, plagiarism cases. Total essays and exams marked = around 170; total plagiarism cases that I worked on = 5. To give each essay a thorough search for plagiarism would quadruple the amount of time to mark them.
The counter-argument to my call for harsh penalties will be: you're going to fail someone who was just ignorant. I think the chances of that are surprisingly small. Students are so much smarter than we usually give them credit for. And those chances can be reduced further by clearly delineating to students the proper use of sources. If we tell them repeatedly what plagiarism is and why it's wrong, then I don't think there's any need to give them 'the benefit of the doubt'. (Again, students who cite sources but don't put things in their own words are not being failed here; we are talking about cutting and pasting from a source, not using quotation marks, and not citing the source or putting it in the biblio.)
Or, perhaps I'm just a big meanie.