[cparedes]

I think the problem isn't inherently the class content (in 'soft' subjects), but how decent the professor is in letting students get by with weak arguments in their papers. I think there's some correlation between how much they actually care about the class, and how tough the class will end up being.

I took an English class in college that was probably one of my toughest classes in college (yes, about as tough as my math classes) - the professor and the TA's tore apart each paper ruthlessly, and judged them at a quality slightly below academic papers. I wrote a paper that was well formatted and had decent grammar in it, but it was summarily torn apart and was given a C on the paper, because my arguments in the paper sucked.

Also, I'd say that Philosophy is also a pretty damn tough subject in the 'soft subjects' of college. You're expected to be nearly as technical as a mathematician in your arguments, except you're writing essays/articles in prose, rather than with terse explanations and symbols. Philosophy took about as much time for me to study for as my math classes.

[arctangent]

I'd agree that Philosophy is a tough subject.

The reason I'd give for this is that in some sense philosophy does have correct answers, if and only if a question is framed within an existing philosophical model.

The point of philosophy (as I see it) is trying to work out which philosophical model is most correct - a bit like a choosing between scientific theories.

Instead of testing against measurable things in the real world, philosophers derive logical outcomes from propositions within linguistic/logical systems and compare them to detect inconsistencies.

Therefore it's fine to ask a question about the consequences of a particular line of thought if a particular philosophical framework is specified, because you can mark a student not only on whether they arrived at an expected answer but also on the quality of their argument (i.e. how they got there).

[sliverstorm]

Would it be wrong to call philosophy "abstract {math|logic}"?

[ezyang]

A large an interesting part of being a philosopher involves logic in its purest sense, but it wouldn't right to claim that this is all philosophy is.

[cafard]

Yes, or at least incomplete.

[jseliger]

The only word missing in your comment is "incentives:" in academia, professors, especially in soft subjects, don't have much of an incentive to grade accurately; they have an incentive to publish "research" that leads to jobs, tenure, and promotion.

If you want professors on the whole to be better and more accurate graders, you have to give them incentives to do so. Which I discuss in more detail here: http://jseliger.com/2011/04/02/grade-inflation-what-grade-in....

[bane]

I have to agree w/r to Philosophy. But I've never considered Philosophy as a "soft" subject in the same way as a history class, English lit class, or some similar class.

Philosophy is every bit as difficult as any advanced Math course I'm familiar with. You learn the rules and axioms of the philosophy, and work within it, and those rules can be as complicated as any other kind of complex logic.

I'm specifically talking about what's considered an undergraduate degree in "English" or "History" or "English History" or some such. I had the fortunate happenstance to go through an undergraduate program that really wanted their engineers to be well rounded. Back in those days we were required to take 3 years of 2 classes a year of English Lit, 2 years of History (a year of Western Civ and a Year of some non-Wester civ), Speech, Communications, etc. for every Engineering Degree and a couple other "soft" courses.

These classes were more annoying than hard. Once you figured out what the teacher was looking for (usually within 2 or 3 short essays), you could write all day on that subject and in that style and get A's. It was literally "programming" for the teacher. Good grades were programs that ran well, bad grades due to grammar or spelling issues were syntax errors, bad grades due to content were run-time.

[FreakLegion]

You're making two mistakes here.

The first is that you're confusing the quality of your (from the sound of it, lower division) undergraduate English program with the quality of English as a discipline. The study of literature, in English or any other language, is, at its core, philosophical.

The second is that no discipline, not even engineering, is immune to students coasting by doing the bare minimum. Criticizing English because you breezed through a couple of lower division surveys is like criticizing computer science because you breezed through a class on HTML.

[quanticle]

The difference is in the level of the bare minimum. Engineering and science have objective, right-or-wrong answers. At the end of the test, either you got the right answer, or you didn't. This imposes a certain minimum level of rigor that, in my experience, is wholly missing from the liberal arts.

Yes, if taught well, the liberal arts can be as rigorous as mathematics. But teaching, say, English, well is whole lot harder than teaching mathematics well. There's just a lot more subtlety that has to be communicated. This means that in practice, there are proportionally fewer good Liberal Arts teachers than there are good math teachers. The quality and rigor of undergraduate liberal arts suffers as a result.

[FreakLegion]

How are you getting from right-or-wrong answers to rigor? That doesn't follow. In fact, at a professional level you could make the argument that a lack of right-or-wrong answers actually calls for a greater degree of rigor. To quote Heidegger:

Precisely from the point of view of science, no field takes precedence over the others, neither nature over history nor vice versa. No one method of dealing with objects dominates the others. Mathematical knowledge is no more rigorous than philological-historical knowledge. It merely has the character of "exactness," which is not the same as rigor. To demand exactness of the study of history goes against the specific rigor of the humanities.

Anyway, anecdotally, four years of shepherding would-be engineers through basic writing requirements was enough for me to conclude that as a group, engineers are singularly terrified by the absence of right-or-wrong answers. For my part, I'm glad to have had a rigorous grounding in the humanities before stumbling into coding.

It's like what Eric Raymond says about Lisp: "Lisp is worth learning for . . . the profound enlightenment experience you will have when you finally get it. That experience will make you a better programmer for the rest of your days." A sufficiently rigorous study of art and philosophy and history won't just show you new things; it will give you new eyes. It's a shame most people (including, of course, many with humanities degrees) miss out on this.

[bane]

I think both your two points here are interesting:

Re: Heidegger: To counter, there's a fairly accepted concept in the sciences that there is a loose hierarchy of the sciences. e.g. Biology is just a subset of Chemistry is just the study of a part of Physics, which is a study of applied Math, which is a subset of Mathematics, which is just a specialization of a subset of Philosophy.

It reminds me of this http://flowingdata.com/2011/06/08/all-roads-lead-to-philosop...

Re: Engineer's proclivity towards right-wrong answers in their work. I agree, and it may be a function of the kind of thinking style that lends itself well to engineering disciplines. The curricula I looked at for Engineers seemed to express an intense desire on the part of the schools to break their engineers out of this kind of mental slavery and to grok softer subjects as valid. Many people who want to become engineers do so because they've mentally dismissed the validity of subjects with non-binary outcomes. I think their is a need to put engineers through a well rounded curriculum that includes exposure to the arts. I think this is very important.

But it means as a side effect that they receive a non-trivial portion of a standard liberal arts education on top of the study of their own field. (In my opinion schools that don't provide this kind of education to their engineers are doing their students a great disservice.)

[_delirium]

I agree there is probably a quality difference in average teaching, but I didn't find much of a difference in the objective/subjective nature (I was a CS major with a philosophy minor).

In the lower-division courses, both CS and most liberal arts have objective answers you can test on, many of which are memorization. Comparison-based sorting can't be done in better than O(n log n); Marx was not present at the Paris Commune (neither was Bakunin, though he was involved in a related uprising in Lyon); most hash-table designs have amortized constant time for most operations; a key difference between Fichte and Kant was that Fichte rejected the notion of a thing-in-itself; etc.

I think it may be true that CS is more likely to actually test on them, in part because many universities expect their liberal-arts courses to double as de-facto "how to write" courses, so they're supposed to have students do essays rather than test them on the subject matter. Imo they would be better structured as having intro-level classes teach subject matter that's tested on in a more rigid style, and then move students to developing their own arguments in essays more towards the 201 level once they've gotten a certain basic amount of factual knowledge under their belt. But that would require teaching intro writing skills somewhere else, and the main alternative, a dedicated "freshman composition" type class, is pretty widely disliked by students and not considered a great success, because it's sort of disembodied writing, "how to write" without any actual subject matter.

Once you got to upper-division courses, most of my philosophy and my CS classes were pretty similar in terms of not really having objective, right-or-wrong answers. There were definitely things you could do that were ridiculous and therefore wrong, but for the most part it involved making a case for something.

A data-mining project, for example: Given a data set, what can you conclude from it? What's your evidence, what are the potential pitfalls with your evidence, how would you present the results?

The methodology is different, but in terms of general approach and the subjectiveness of grading, that felt very similar to me to a philosophy course project, one of which was: develop and argue a case for or against the possibility that machines could produce "creative" output.

In both those examples it has less to do with there really being a "right" answer than with being sufficiently fluent with the tools of the domain to build and coherently present a supportable case, while avoiding doing anything that's clearly "wrong", like misusing statistics or using examples that don't logically support your point.

[bane]

A data-mining project, for example: Given a data set, what can you conclude from it? What's your evidence, what are the potential pitfalls with your evidence, how would you present the results? The methodology is different, but in terms of general approach and the subjectiveness of grading, that felt very similar to me to a philosophy course project, one of which was: develop and argue a case for or against the possibility that machines could produce "creative" output.

I actually think that supports my idea here -- that a competently trained engineer should be able to operate competently in a traditionally liberal arts field -- because a competently trained engineer has to learn both skill sets. Upper level courses like Data Mining require both the engineering bits, and the art bits of the liberal arts.

Interpreting the results of a Data Mining algorithm requires a similar critical analysis and writeup of any upper level Lit coursework. The demonstration of how the student wrote their analysis is as important as their conclusions.

This is different than say "write a paper detailing the differences and similarities between decision trees and neural networks", which is akin to "write a paper about the goals and purposes of the Paris Commune -- and why it failed".

The differences then between an engineer and a English major, in the field of English, is not a matter of skill, but a matter of exposure. An engineer is less likely to have studied in depth the brief life of the Paris Commune, but should be able to competently write about it once they know the material. What makes Engineering hard as a discipline is that the Engineer must not only have memorized the material, and be able to competently write about it, they must also apply it and show it working. That 3rd bit is why people drop from engineering programs and where the right/wrong objective evaluation brutalizes the unprepared student who was expecting subjective evaluation of their work.

[FreakLegion]

You're right -- most freshman composition programs simply aren't very good. There's a distressing amount of politicking going on behind the scenes there. I actually used to sneak interesting material in by lying to the department about what I was teaching. But even then, the fact that students enrolled because it was a requirement, and not because of what I was teaching, made it difficult to find material with a sufficiently broad appeal. You have to be a really good teacher, and in some ways a really bad employee, to teach composition well.

[bane]

"The first is that you're confusing the quality of your (from the sound of it, lower division) undergraduate English program with the quality of English as a discipline. The study of literature, in English or any other language, is, at its core, philosophical."

I honestly hope that you can illuminate me -- the corpus of books available in the language are the same for both you an I. Before school, I toyed around with being an English major and a survey of programs at the schools I was interested in didn't really seem to offer course loads that would have provided me with any particular insight except maybe an extended survey of Milton's work.

Years later, based on my anecdotal experience, I find it obvious that my school mates who went on to major in liberal arts can't find work that pays as well as engineering. Any decent engineer could, more or less, do their jobs. Supply far outstrips demand and the training to become an Engineer requires a superset of the skills required to graduate with a B.A. in English.

Again, based on my anecdotal and singular experience, suppose a massive nuke went off in the atmosphere and the EMP wiped out all electrical equipment on the planet. I have no doubt that any competent engineer that went through a school with a reasonably well rounded curriculum would find work as a copywriter for some periodical someplace. A week or two to the learn the style and content the editor wants and off you go. It literally isn't rocket science. If they can effectively communicate in writing, they can participate at a reasonably high level in English as a discipline. The opposite is not true. Very few English majors could "fall" into an engineering job and be up and running in a couple of weeks. Engineering is simply harder and sticking it to a B.S. in <insert Engineering discipline> requires the student to almost learn a superset of the skills needed to get a B.A. in <insert English discipline>.

A spot check of programs at highly ranked liberal arts schools (I'm avoiding well known engineering schools on purpose here) shows coursework necessary for a B.A. in English Lit involves classes in "American Lit: Civil War to Present", "Chaucer, Shakespeare and Milton", "Myth, Symbol and Allusion in Literature" (yes, actual title), "British Literature 1785 to Present", "Latin American Literature (In Translation)", "Film, Media and History" and electives that include "pick a major author (approved by your Faculty Advisor)", "any writing course", "any historical period including modern writing", "Film" (or as I like to call it "watch some movies and write a 5 paragraph essay on what you just watched". In other words, everything that every engineer at my school either had to do, or just did as part of their normal business in other coursework. Not exactly mind-bending stuff.

Here's Amherst College's (I picked them since they are one of the top ranked Liberal Arts schools in the country) actual program https://www.amherst.edu/academiclife/departments/english/maj...

"Majoring in English requires the completion of 10 courses offered or approved by the Department. Students are encouraged to explore the Department’s wide range of offerings in literature, film, and culture. Rather than prescribe any particular route through its curriculum, the Department helps its students develop their own interests and questions, which are then recorded in the students' concentration statements."

Yes that's right, 10 whole courses. By way of comparison, I had to do 50% of Amherst's English major for my Engineering degree just as a minimum requirement. Had I stayed on an extra semester I could have knocked off the equivalent of the entire program had my school offered Amherst's program.

You must take One Level I course, Three Level II courses, One Level IV course, and just to make it extra hard, a course addressing pre-1800 material (which usually really means 1500-1800, heaven forbid anybody study Beowulf). In other words, Intro, three sophomore levels, one proper senior level, and then Chaucer, Shakespeare and Milton. Not exactly stunning stuff. Brilliantly, I didn't see a single school that had a course about the century of Science Fiction and Fantasy literature that fills bookshelves these days.

Yeah there's some levers the schools can pull to simulate hardness: pile on more readings, assignments in longer essay forms, period lit, translations. But at some point it just becomes busy work. Either you can read a work and write an essay to the assignment or you can't. Figuring out the particular philosophy the instructor wants you to emphasize in their class is an added twist, but it isn't like solving Fermat's last theorem.

By comparison, a random state school I looked at offered a sample schedule for CS students that required an English course per year through their Junior year, A year of Western Civ, Social and Behavioral Sciences, Fine Arts, Something called Global Understanding, Ethics and Law, 16 courses in CS, 4 in Mathematics, 2 in nonrelated engineering such as EE or Civil, 3 semesters of Natural Sciences with labs and must require calculus in the coursework...no Earth/Environmental Sciences for Arts majors, and on top of that, just because CS majors need extra help, a public speaking course and one more humanities course. In the course of a single week you might go from writing a paper on Chaucer, preparing a speech, finishing your chemistry lab, writing an operating system and then back to a paper on Charlemagne followed by homework in Statistics. No wonder people drop out of Engineering disciplines. This kind of curriculum is unbelievable hard.

A specific example a la Amherst, here's Georgia Tech's core CS requirements (this doesn't even include the actual CS courses)

http://www.cc.gatech.edu/future/undergraduates/bscs/corereq

4 Humanities Courses, 4 Social Sciences, 3 Lab Sciences, 5 Maths, 4 other courses, then you can get to the requirements for the major.

So at Georgia tech, just to get to the part of the your degree that you want to take, you have to accomplish 40% of Amherst's English major and a bunch of other stuff.

I agree that there are some elements of advanced literary analysis and writing that my figurative competent engineer might struggle with because of lack of domain knowledge. I certainly couldn't write an extended critique on the use of monologues in English translations of 18th Century Portuguese Enlightenment critiques off the top of my head -- or write an extended Avant Garde piece without falling into obvious uses of Absurdism. But I doubt many undergrad English Lit majors could do the same (I have yet to meet any myself) since the coursework at most schools I looked at didn't touch these subjects at all. I couldn't find a single undergrad English course in my survey of B.A. programs that looked even the slightest bit more daunting than what I had to take for my engineering program. A class on Milton? Pish Posh! A class on North African Poetry? No sweat! Depression Era Authors? Puffincakes!

"The second is that no discipline, not even engineering, is immune to students coasting by doing the bare minimum. Criticizing English because you breezed through a couple of lower division surveys is like criticizing computer science because you breezed through a class on HTML."

It's true, there are different levels of programs. While writing this I ran across a couple questionable programs that probably wouldn't put out a very good engineer. But something to keep in mind is that in most of the engineering programs I'm talking about, the minimum requires means that you got things undoubtedly correct. You got the right answers in the math exams, your programs ran and gave the expected output (and passed the anti-cheat detection software), you built the clocked full-adder correctly and it lit up the right number of lights, you derived Newton's formulas from your lab experiments correctly, or measured the pH of the solution without contaminating the sample. You did all this while also taking all the Reading, Writing, Memorizing and Speaking classes your school demanded of you.

You can't just fill up a paper with meaningless koan vomit, half of it copied from the Internet and get by with a C. If your full adder tells you that 2+3 = purple and your labwork shows that the acceleration of gravity near Earth's surface is 73.9 m/s^2 you get an F.

A grade other than an A or F doesn't mean you successfully argued the point to the professor that your code wasn't meant to run properly as a statement about the absurdity of Women's roles in the Elizabethan Nobility and your teacher just tossed you a bone to get you to stop harassing him on his personal phone number at 3am. It means that it still ran, but failed to provide the expected output for 1 out of a billion possible inputs. It's wrong and buggy, but you don't fail the course because in the real world QA and proper automated tests should catch that anyways.

I remember one course in particular, over the semester my 3 person team wrote a minimal multi-tasking operating system and software stack that included a pretty full user space, a text editor, system level IM and email, an ftp client and server, etc. and got a C+ because we never were able to squash the one bug that didn't allow the system to host itself, even after devoting 2 solid months of 90 hour weeks to it. Oh, and we had to write a paper, build a website, an instruction guide, and about 100 other pages of material for the system. That bug was the only thing that we didn't get perfect marks on and that was all it took, goodbye 4.0.

Back to the main question then, why do people drop out of engineering? Because it's goddamn hard. People will fall into other majors that are simply put easier.

[wtvanhest]

Is it possible that soft subjects are easier not because they are soft, but because weak competition self selects in to those classes?

[shareme]

come on..that was my senior English in high school 20 years ago..and yu call that hard?