Many generations of students have had certain grammar “truths” drilled into their little heads. One is the “myth” that infinitives can’t be split. But today we’re going to discuss the myth that sentences can’t start with conjunctions.
(Actually, whether teachers do indeed prohibit conjunctive beginnings seems to be almost as much a myth as the prohibition itself. But more on that later.)
Conjunctions, which join things in grammar, come in three flavors: Coordinating, which link things directly (“and,” “but,” “for,” “nor,” “or,” “so,” etc.,); correlative, which are used in pairs and connect parallel things (“both and,” “either or,” “neither or,” “not only but also,” “whether or,” etc.); and subordinating, which usually join clauses to other, nonparallel, parts of the sentence (“after,” “although,” “before,” “if,” “since, “that,” “while,” etc.) Note that many subordinating conjunctions have multiple roles in grammar: “After,” for example, is also a preposition; “that” can be a relative pronoun; and many subordinating conjunctions are adverbs as well.
That’s a lot to remember, which may cause some of the confusion: If you can’t start a sentence with “Neither he nor I got any dessert,” options for starting sentences become extremely limited. Which is one reason any prohibition is silly: You end up making most sentences subject-verb-object, which can get pretty boring, not to mention tortured: (“Dessert was given to neither he nor me.”)
The usual suspects, though, are the coordinating conjunctions. So we’ll make it easier:
It’s perfectly OK to start a sentence with “and,” “but,” “or,” and all of those conjunctions. The Bible does it; the most persnickety writers do it; grammar authorities do it. Even going back to early Fowler (A Dictionary of Modern English Usage), the prohibition on conjunctions was being dismissed. “That it is a solecism to begin a sentence with and is a faintly lingering superstition,” the second edition says, echoing the first edition.
The American Heritage Dictionary notes that “this rule has been ridiculed by grammarians for decades, and the stricture has been ignored by writers from Shakespeare to Joyce Carol Oates.” (This is from the Fourth Edition; the Fifth is in the mail.)
But whose rule is it, anyway? Of the dozen or so grammar books intended for grammar schools that we consulted, not one bars conjunctions at the starts of sentences. No reliable grammar website bars them, either.
So who’s prohibiting it, and why? Among the best explanations comes from the Language Log, where the linguist Arnold Zwicky calls the “rule” “No Initial Coordinators,” or “NICs.” Zwicky posits that, because children often overuse “and” in spoken language (“And then we had ice cream. And then we had recess.”), teachers decided “‘If they do it too much, they should be told not to do it at all’, and NIC, a blanket proscription, was born. Probably in elementary schools, from which it would have diffused to secondary schools and beyond. And now the zombie lurches on.”
And as he notes, some versions of Microsoft Word’s grammar checker will flag a conjunction at the start of a sentence (another reason to not use it). Where does Microsoft come up with that rule? Its grammar checker is “fully developed and owned by Microsoft,” the company says. So there.

"... to neither he nor me?" Please, let's not muddy the water with further grammatical errors.
#1 Posted by Linda Thurston, CJR on Wed 9 Nov 2011 at 12:39 PM
At Penn State, some of the worst operators are being cleared out. At UPenn, Mark Liberman is still getting away with it.
His work at Language Log is a drag on progress in linguistics in America. As an example, IELTS has come around as an important subject again, this time in Australia. When Mark took up this subject on Language Log, he got so unfocused he attacked himself viciously. An entirely appropriate reaction. Instead of choosing a sample IELTS manual and comparing it with the COBUILD Intermediate English Grammar, he talked nonsense and went in such violent circles that he fell down. Hard.
As a teacher, if comments below are any indication, he is atrocious. He refuses to listen to advice, rejecting it with contempt.
His ideas about how to teach the sound systems of English meaningfully are absurd. He has next to no understanding of how to link corpus dictionaries and grammars in teaching programs so as to prepare students for literature.
The effective result of his hateful behavior is to disadvantage learners of English, especially second language learners.
He has been the subject of a complaint to the UPenn administration, but obviously impunity is the norm.
Mark Liberman UPenn Linguistics. Fall 2011 Linguistics 001: Introduction to Linguistics COGS 501: Mathematical Foundations of Language and Communication Sciences.
Comments at ratemyprofessors.com:
Brilliant in his field, but he can't communicate it in his lecture classes (ironic, as he is a linguistics prof). He reads straight off of his very text-heavy notes (which are online, meaning that you can learn everything without once going to lecture) and could use some work in public speaking. The spring professor's supposed to be much better.
Terribly boring. RUINED my interest in linguistics. It's a shame because I would have been a linguistics major (instead of an English major) if I had a better Ling 1 professor. I didn't rediscover linguistics (with a better prof) until my senior year. One of the worst profs at Penn.
SO BORING
Name: Mark Liberman School: University of Pennsylvania Location: Philadelphia, PA Department: Languages 1.7 OVERALL QUALITY 1.7 HELPFULNESS 1.7 CLARITY 3.0 EASINESS 0 HOTTNESS.
#2 Posted by Clayton Burns, CJR on Wed 9 Nov 2011 at 07:39 PM
The best program for study of these issues in American schools would be repeated and minute focus on the COBUILD Intermediate English Grammar and COBUILD English Grammar. Study of sentences through the CD of the excellent Longman Dictionary of Contemporary English is also valuable.
The Language Log approach to English grammar is obsolete and unimaginative. We need to be able to diagnose a pathology such as misunderstanding of the depth of the resources of the past and construct systems so that we will be able to fit elements into sentences.
The most interesting sentence type is this counterfactual: "If you had wanted to let in the fresh air and keep out the flies, the best thing to do would have been to fix up a screen."
"If you had fixed up a screen, that would have been the best thing to do so as to let in the fresh air and keep out the flies."
"The best thing to do is to fix up a screen so as to let in the fresh air and keep out the flies."
Once you have mastered such transformations, it gets easier. The reason that NYT's On Language and Language Log have failed is the lack of depth on the fundamentals of language. Especially in grammar.
#3 Posted by Clayton Burns, CJR on Wed 9 Nov 2011 at 08:03 PM
The Daily Pennsylvanian is an honest paper that publishes reasonable comment. Mark Liberman of UPenn, unfortunately, is unfit to be a professor. He is a hateful and an incompetent person.
Even though this evening I posted a friendly comment at Language Log, containing my 60 verb elements of the past, invaluable information for teachers of English, everything was deleted again.
Mark Liberman of UPenn Linguistics hates knowledge.
#4 Posted by Clayton Burns, CJR on Wed 9 Nov 2011 at 09:57 PM