P.

Schlenker - Ling 1 - Introduction to the Study of Language UCLA, Winter 2006

Introduction to the Study of Language - Lecture Notes 1A

Rules of Language:
Description vs. Prescription

[Last update: 01/09/06]

☞ Goal: To know a language is to have in principle the ability to utter and understand infinitely many new
sentences. How is this possible? The key is that speakers know (a finite number of) rules, which can be applied
repeatedly to produce an infinite number of sentences. All spoken language is in this sense governed by rules;
in this respect there is no difference between what is considered 'good English' and what is considered 'bad
English' - they each follow rules, though they may be different ones. We will distinguish two uses of the
notion of 'rule': for prescriptive vs. for descriptive purposes. Prescriptive rules are intended to teach people
how they should speak or write according to some pre-determined (arbitrary) standard. They are of dubious
origin, have no linguistic justification, and have no relevance for the linguist, who is solely interested in
describing and understanding the rules that speakers do in fact follow (=descriptive rules).

1 The Miracle of Language: 'Infinite Use of Finite Means'

1.1 Any speaker can in principle construct an infinite number of sentences.

-There is no limit to the complexity of the sentences that we may in principle utter or understand:
(1) a. John is asleep
b. The President is asleep
c. The chairman of the Linguistics Department is asleep
(2) a. The rightmost person in the first row is asleep.
b. The person immediately to the left of the rightmost person in the first row is asleep
c. The person behind the person immediately to the left of the rightmost person in the first row is asleep

(3) a. John is asleep

b. Mary noticed that John is asleep
c. Nobody cares that Mary noticed that John is asleep
d. Sam knows that nobody cares that Mary noticed that John is asleep

-Note that if we modify the order of words in any of these sentences, or omit some of the words, we typically
obtain sentences that are 'odd' or 'sound weird'. For this reason they will henceforth be preceded by a star: * As
we will see, the fact that a sentence is 'odd' or 'sounds weird' indicates that a rule of the language has been
violated.
(4) a. *asleep is John
b. *President the is asleep
(5) a. *Rightmost person the in the first row is asleep.
b. *The to the left of the rightmost person in the first row is asleep
P. Schlenker - Ling 1 - Introduction to the Study of Language UCLA, Winter 2006 2

1.2 Speakers know (a finite number of) rules

-The puzzle, then, is this: how can human beings, who are finite creatures, have knowledge of an infinite
number of sentences? The answer is: because they know rules, which they may apply repeatedly to form any
number of new sentences. In this respect our knowledge of language is similar to our knowledge of integers. In
principle there is no limit to the number of even integers that you may enumerate: 2, 4, 6, 8, etc. In this sense
you know an infinite number of even integers. But of course you did not memorize them all! You know a rule
that allows you to produce 'new' even integers from old ones. In this case the rule is: for any even integer E,
E+2 is another even integer.

-What is the longest sentence of English? Answer: there is none! For any sentence that you care to choose, you
may construct a longer sentence by embedding it under 'Nobody cares that __'.

Examples:

(6) a. John is asleep ➔ Nobody cares that John is asleep

b. Mary noticed that John is asleep ➔ Nobody cares that Mary noticed that John is asleep

This is our first example of a rule of English:

A simple rule: If S is a sentence of English, Nobody cares that S is a sentence of English.

Of course as stated this rule is very crude, but it does give us an idea of how we may have knowledge of a
potentially infinite number of sentences. Notice that this rule can be iterated (=repeated) any number of times
to produce an arbitrary number of new sentences:

(7) a. John is asleep

b. Nobody cares that John is asleep
c. Nobody cares that nobody cares that John is asleep
d. Nobody cares that nobody cares that nobody cares that John is asleep
e. etc.
[Note that if all you have memorized is (a) the sentence 'John is asleep', and (b) the simple rule described
above, you will still be able to produce a potentially infinite number of sentences, simply by repeating the
procedure whose results are illustrated in (7)]
P. Schlenker - Ling 1 - Introduction to the Study of Language UCLA, Winter 2006 3

2 Prescription vs. Description

When we speak of rules of grammar, we often mean prescriptive rules, i.e. rules that are intended to tell
people how they should speak or write according to some pre-established (arbitrary) standard. Prescriptive
rules are of dubious origin and have no linguistic justification. The linguist is solely interested in
understanding descriptive rules, i.e. rules that govern the way in which people actually do speak. Every spoken
language is governed by rules in this sense. This does not mean that every speaker of English follows exactly
the same rules: English has a number different dialects, which are equally valuable but are nonetheless distinct.

2.1 Some Prescriptive Rules of English [C. Phillips, University of Maryland]

♦ Examples
(8) Don’t split infinitives!
a. Do not say: I wanted to carefully explain to her why the decision was made.
b. Say: I wanted to explain to her carefully why the decision was made.
(9) Don’t use double negation!
a. Do not say: I didn't do nothing
b. Say: I didn't do anything
(10) Don’t end a sentence with a preposition!
a. Do not say: A preposition is not a good word to end a sentence with
b. Say: A preposition is not a good word with which to end a sentence.
(11) Don’t use who in place of whom!
a. Do not say: Who did you talk to?
b. Say: Whom did you talk to?
Some of these rules stem from an attempt to make English look like Latin. Thus in Latin an infinitive, being a
single word, could never be split. But of course from this it does not follow that the same should hold of
English, where 'to explain' is made of two words, not one.

In any event, the linguist as a scientist has nothing to say about prescription. We will attempt to describe and
understand the rules that speakers do in fact follow; we are interested in how people speak, not in how they
should speak. Accordingly we distinguish between descriptive grammar and prescriptive grammar:

Descriptive grammar has as its goal to describe what the native speakers of a language do (verbally)
when they speak their language (the meaning of the word “grammar” as used in this course).
Prescriptive grammar categorizes certain language uses as acceptable or unacceptable according to a
standard form of the language (the meaning of “grammar” normally intended in English classes).
P. Schlenker - Ling 1 - Introduction to the Study of Language UCLA, Winter 2006 4

An example:
Use of slow vs. slowly and similar pairs of adjectives vs. adverbs (“adjective form” here refers to the word
without -ly, “adverb form” refers to the word with -ly):
Descriptive rule Prescriptive rule

There is a certain overlap between the adjective and “Use as an adjective a word which qualifies a noun. Use as an adverb a
adverb classes, e.g. the adjective form slow may be word which qualifies a verb.” (Greever & Jones, The Century Collegiate
used as either adjective or adverb. However, when Handbook, 1924)
the adjective form is used as an adverb, it must
follow the verb; only the adverb form is allowed
preceding a verb. (Adapted from Quirk, et al., A
Comprehensive Grammar of the English Language,
1985)
You drive too slow! (conforms to descriptive account but violates prescriptive rule)
You drive too slowly! (conforms to both descriptive and prescriptive accounts)
He slowly negotiated the curves. (conforms to both descriptive and prescriptive accounts)
*He slow negotiated the curves. (violates to both descriptive and prescriptive accounts)

♦ Relativity of Prescriptive Rules

What is considered grammatically proper depends on historical circumstances that have nothing to do with
purely linguistic or logical considerations. For instance, in contemporary French double negation is considered
to be 'proper', while single negation is considered 'sloppy' - the opposite pattern from the one we find in
English:
(12) Contemporary French
a. Il ne mange rien Prestige Dialect
He NOT eats nothing
'He doesn't eat anything'
b. Il mange rien Spoken Language
He eats nothing
'He doesn't eat anything'
'Double negation', or 'negative agreement' [also called 'negative concord] is a feature of BEV [Black English
Vernacular], as well as other varieties of English.

2.2 All spoken language is governed by rules!

What is considered 'sloppy' speech turns out to be governed by systematic rules. Here is one extreme example
- a highly improper term which, in slang, can be inserted inside words, apparently in an arbitrary fashion:
♦ Example 1: fuckin-insertion
(13) a. fan-fuckin-tastic
b. abso-fucking-lutely
c. Phila-fuckin-delphia
d. Kalama-fuckin-zoo
e. Pennsyl-fuckin-vania
P. Schlenker - Ling 1 - Introduction to the Study of Language UCLA, Winter 2006 5

As it turns out, the process is systematic, as witnessed by the fact that an odd result is obtained when the term
in question is inserted in different positions:

(14) a. *fanta-fuckin-stic
b. *absolute-fuckin-ly
c. *Kala-fuckin-mazoo
d. *Penn-fuckin-sylvania

In fact, there is a systematic rule underlying the insertion of this term: as a first approximation, it may only be
inserted right before a stressed syllable, represented in bold below:

(15) a. fanTAstic ➔ fan-fuckin-tastic

b. absoLUTEly ➔ abso-fucking-lutely
c. PhilaDELphia ➔ Phila-fuckin-delphia

Descriptive Rule: 'fuckin' can be inserted only before a stressed syllable.

♦ Example 2: missing 'be' in Black English Vernacular (BEV; you will also encounter the term 'African-
American English Vernacular', or AAEV)
[W. Labov, 'The Case of the Missing Copula', in Gleitman & Liberman (eds), Language, 1995]

Note: The examples described below are by no means illustrative of the speech of all African-Americans, just
of one dialect among others (these examples are reproduced because they were the object of an influential
study by the sociolinguist Willian Labov).

(16) a. She the first one started us off (Dolly R., 35)
b. He fast in everything he do (M, 16)
c. Michael Washington out here selli' his rocks (F, 14, East Palo Alto)
(17) a. Boot always comin' over my house to eat, to ax for food. (M, 10, South Harlem)
b. He just feel like he gettin' cripple up from arthritis (F, 48, North Carolina)
c. Y'all got her started now, she fixin' to give y'all a lecture! (F, 14, East Palo Alto)
We now consider two hypotheses about these data.

Hypothesis 1. 'Be' can be freely omitted in BEV

Hypothesis 2. 'Be' is not omitted in BEV, but it is phonologically reduced - a more extreme form of the
phenomenon found in standard varieties of English: John is nice ➔ John's nice
P. Schlenker - Ling 1 - Introduction to the Study of Language UCLA, Winter 2006 6

Argument against Hypothesis 1: In some contexts, be cannot be omitted in AAVE

(18) Imperative
a. Be cool, brothers! [M, 15, South Harlem]
b. Don't be messin' with my old lady! [M, 16, South Harlem]
(19) Emphasis
a. Allah is god [M, 16, South Harlem]
b. He is a expert [M, 12, South Harlem]
(20) Yes-no questions
Is he dead? is he dead? Count the bullet holds in his motherfucking head. [M, 16, South Harlem]

Argument in favor of Hypothesis 2: Be can 'disappear' in BEV in exactly those contexts that permit elision of
is to 's, am to 'm and are to 're in English.

Other English dialects Black English Vernacular (BEV)

a. *He's as nice as he says he's *He's as nice as he says he
b. *How beautiful you're! *How beautiful you!
c. Are you going? *I'm Are you going? *I.
d. *Here I'm. *Here I.

Although we have not explained what the precise rules are, it is clear that omission of be in BEV is governed
by strict rules.

Conclusion: (i) Any spoken language is governed by rules.

(ii) By knowing a finite number of rules, speakers can in principle utter or understand an
infinite number of new sentences.
(iii) The linguist is interested in descriptive rule, not in prescriptive rules.