Lexical Semantics

Final Exam
Institut für Sprache und Information
Heinrich Heine Universität
Instructor: Prerna Nadathur

January 20, 2020

[minor update/correction: January 29]

Submission deadline: February 4, 2020, 23:59pm

This is a hard deadline. Late exams will ONLY be accepted in case of an

emergency.
You may submit your assignment by email to [Link]@[Link], or turn in
a paper copy to the Lexical Semantics box in the Linguistics department office ([Link]).
Please alert me by email if you hand in a paper copy. Your answers should be in
English unless you have made prior arrangements with me.

Reminders:
• Your final exam solutions should be typed.
• You may NOT work with any other students on the final exam.
• You may refer to any of the class handouts, readings, or previous assignments.
• You may consult any additional resources (web pages, journal articles, books) that
you wish, as long as you cite these materials.
• You will NOT be penalised for mistakes related to English grammar.

Additional instructions:
The questions on this exam are very similar to the questions on the homework assign-
ments, and will be marked according to a similar scheme. You must receive 60% of the
available points in order to pass the exam. There are 150 available points.
The exam is divided into sections, each of which covers different topics from the
course. Note that questions may require you to bring together more than one topic from
different parts of the course. For example, while the questions in Part III deal mainly
with lexical aspect, you will also need to think about semantic roles and gradability.
You must complete the specified number of questions from each section. If you com-
plete more than the required number of questions, your score will be calculated based
on those questions for which you received the most points (in other words, doing more
questions can only improve your score).

1
Part I: Inference and implication types.
Answer two of the following three questions. [25 points each]

Question 1: Types of implication

Spell out three implications of each of the following sentences. For each implication,
specify what type of implication it is (entailment, presupposition, or Gricean implicature)
and give a brief explanation for your designation. Your explanation should include which
words, morphemes, or phrases are responsible for any given implication. One sentence is
marked as ambiguous. For that sentence, you must identify one source of ambiguity and
spell out the possible readings. Then indicate which of the three implications you provide
are shared by all the readings, and which ones are specific to a particular reading.

(1) a. Mary forgot that John was happy that Sue had the opportunity to participate
in the race.
b. It is significant that John failed to remember to straighten the tablecloth.
c. If they managed to open the entrance to the cave again, they had shelter for
the night. ambiguous
d. It doesn’t bother Mary that John criticized Bill for writing the letter.
e. Mary regretted that John didn’t have the sense to shorten the rope.

Question 2: Factives and implicatives

Classify the verbs and constructions marked in bold in the following sentences as factive,
implicative, or neither. Explain your answers. Your explanation should include tests for
different kinds of inferences.

(2) a. A full investigation revealed that the company’s CEO was stealing money.
b. Jeremy was afraid to go skiing for the first time.
c. Morgan made an attempt to solve the riddle.
d. Masha’s boss took credit for her work on the project.
e. Remy took the opportunity to finish his short story.
f. Nima was lucky to get a ticket to the concert.
g. It is strange that no one came to school today.
h. Tammo happened to be in Boston on the day of the marathon.

2
Question 3: Implication propagation
(a) The sentence in (3) contains four verbs, and each verb is responsible for at least
one implication, due to its primary meaning or to a presupposition. Spell out four
inferences associated with (3), one for each verb. In each case, specify what kind of
inference it is (entailment, presupposition, etc), and identify the responsible verb
as factive, implicative, or neither.

(3) Mary regrets that Ed did not forget to force Dave to leave Athens.

(b) Provide the implicational signatures of the relevant lexical items (the relevant verbs
and not) and show step by step how the projection of implications that are based
on the unembedded clauses can be calculated.

Part II: Ambiguity and polysemy.

Answer one of the following two questions. [20 points each]

Question 4: Adjective polysemy

Consider the adjective empty in the examples in (4):

(4) a. The road was empty. d. The house was empty.

b. The car was empty. e. The classroom was empty.
c. The vase was empty.

(a) Spell out potential implications of all of the examples in (4). The implications
should be based on the adjective empty and can be of the form there were no X in
N, where N is the head noun of the subject noun phrase.
(b) Can you identify a unique sense for the adjective empty in all the examples in (4)?
If not, how many senses are needed, and how can they be characterized?

Question 5: Ambiguities with almost

(a) Identify two distinct readings of (5) and spell out what they are. You can do so
either by describing situations with incompatible properties which can be truthfully
described by (5), or by indicating the implications of (5) on each reading.

(5) They almost emptied the tank.

(b) Provide a decompositional representation of the verb empty. Does this decomposi-
tion explain how the two readings you identified in part (a) arise?

3
Part III: Lexical aspect and aspectual class features
Answer one of the following two questions. [30 points each]

Question 6: Aspect and agentivity

Perhaps because states by their very nature involve no change, descriptions of states
typically do not involve an Agent semantic role. As a consequence, there is some debate
over whether certain linguistic phenomena are best explained in terms of stativity or
agentivity. In this question, you will be asked to examine one such phenomenon and
decide how best to characterize it.
Contrasts like those in (6) and (7) have led some researchers to claim that stative
predicates cannot be found in what are known as “bare infinitival” sentential complements
to the verb see – the type of sentential complement found in (6) and (7).1

(6) a. *Kim saw Lee know French.

b. *Kim saw Lee resemble Pat.
c. *Kim saw Lee be tall.
(7) a. Kim saw Lee skate on the pond.
b. Kim saw Lee carve a whistle.
c. Kim saw Lee miss the target.
d. Kim saw Lee wink at Terry.

In this question, you will assess whether this claim is an accurate characterization of the
restrictions on the complement of see. That is, the goal is to determine whether the
restrictions on the sentential complement are aspectual in nature, as the data in (6) and
(7) suggest, or whether they can receive an alternative explanation in terms of agentivity.

(a) Explain why the examples in (6) and (7) appear to confirm the claim that the
sentential complements are not stative. Be sure to give independent evidence for the
aspectual classification of each sentential complement in (6) and (7) as it occurs in
isolation (i.e., Lee knows French, Lee skates on the pond, etc.) by making reference
to appropriate aspectual diagnostics.

(b) As mentioned above, certain “contexts” that are claimed to be sensitive to stativity
have instead turned out to be sensitive to agentivity. Could the pattern of data
in (6) and (7) be attributed to the (non)agentivity of the subject of the sentential
complement rather than to the sentential complement’s stativity? Explain, giving
1
Note: The verb see also takes other types of sentential complements, such as the that-complement
in Kim saw that Lee had built a tree house. For this question, you should consider only bare infinitival
complements, which are those where the verb in the complement does not agree for person and number
with the subject – e.g., the infinitival form know in (9a), instead of the 3rd person singular form knows,
which would agree with Lee.

4
 independent evidence concerning whether the subject of each of the sentential com-
plements in (6) and (7) (i.e., Lee) is or is not an agent when these complements
occur in isolation by making reference to an appropriate diagnostic.

(c) Explain how the data in (8) can be used to settle the matter of whether the sentential
complement of see is restricted to non-stative sentences or to sentences with an
agent argument. Be sure to indicate which way the matter is resolved! Again,
give independent evidence for the aspectual classification and (non-)agentivity of
all relevant sentential complements.

(8) a. *Clara saw the Golden Gate Bridge be long.

b. Clara saw the boulder roll into the stream.
c. Clara saw the aspen leaves flutter in the wind.
d. Clara saw the rising water reach the window sill.
e. Clara saw the light flash.

Question 7: Verbs derived from gradable adjectives

English verbs derived from gradable adjectives by the addition of the suffix -en, such
as broaden, deepen, flatten, shorten, straighten, sweeten, thicken, widen, or by a zero
suffix, such as clean, cool, dim, dry, narrow, steady, warm, have long intrigued linguists
investigating lexical aspect. A variety of factors affect the aspectual classification of
sentences with such verbs and some of these will be investigated in this question.

(a) Deadjectival verbs do not display uniform behavior with respect to the various
tests for (a)telicity. For example, some, including dim, deepen, shorten, thicken,
and warm appear to be atelic with respect to the progressive-implies-perfect test
(diagnostic test 5, from the top of p.8 of the Week 11 handout), while others,
including dry, empty, flatten, steady, and straighten, appear to be telic with respect
to this test.

(9) a. The cook is thickening the sauce.

→ The cook has thickened the sauce.
b. The stagehand is dimming the lights.
→ The stagehand has dimmed the lights.
c. The workers are deepening the trench.
→ The workers have deepened the trench.
(10) a. The smith is flattening the metal.
6→ The smith has flattened the metal.
b. The sun is drying the dress.
6→ The sun has dried the dress.
c. Smith is emptying the bucket.
6→ Smith has emptied the bucket.

5
 In fact, the telicity of a sentence with a deadjectival verb can be predicted from a
property of its base gradable adjective. Identify this property and explain why it
should have the effect it does on the telicity of a sentence with the related verb.

(b) The deadjectival verbs that are atelic, according to the results of the progressive-
implies-perfect test, may appear with measure phrases, such as the italicized phrases
in (11); these phrases are sometimes introduced with the preposition by. What ef-
fect, if any, does the addition of a measure phrase have on the aspectual classification
of sentences with atelic deadjectival verbs and why?

(11) a. Lee shortened the speech by 5 minutes.

b. The chemist warmed the mixture by 15 degrees.
c. The engineer deepened the well by three feet.

(c) We have observed that verbs such as eat are telic when their object is a singular
count noun, but atelic when it is a mass noun. In this question we have seen that the
telicity of certain deadjectival verbs depends on the nature of the base adjective.
Explain, in your own words, why adjectives like flatten and empty pattern with
singular count nouns in giving rise to telic sentences, while adjectives like deep and
thick pattern like mass nouns in giving rise to atelic sentences? To answer this
question you need to think conceptually.

Part IV: Denominal verbs and the manner/result dichotomy

Answer the following question. [30 points]

Question 8: Verbs of removal

English has a large set of verbs that could be labeled “verbs of removal” since they
describe an agent removing stuff from a location; several of these verbs are listed in (12).
This characterization of these verbs receives support from the fact that they are all found
in the syntactic frame “NP V NP from/off of/out of NP”, as in (13a); in this frame
the location is indicated by prepositions that indicate sources – the location an entity
moves from – with verbs of motion (e.g., Alex jumped from/off of/out of the bed ). Verbs
of removal are also all found in the frame “NP V NP”, as in (13b); in this frame, the
location follows the verb and the stuff is left unexpressed.

(12) clear, rinse, swab, sponge, sweep, empty, shovel, wipe, vacuum, clean, mop, scrub,
rake, scour, hose, wring.
(13) a. Robin cleared/shoveled the snow from/off of the sidewalk.
b. Robin cleared/shoveled the sidewalk.

6
 (a) A closer look at the list of verbs of removal suggest that this class of verbs, like
many others, can be split into a subclass of means/manner verbs and a subclass
of result verbs. Which verbs in (12) are means/manner verbs and which are result
verbs? You need not explicitly justify your classification of every verb in (12) as a
means/manner verb or as a result verb, but you should do this for one verb of each
type. Justifying your classification means explaining why the meanings of the two
verbs you choose conform to the general kind of meaning of verbs of the relevant
type; see class handout from Week 13.
(b) The list of verbs of removal contains three verbs derived from adjectives, clean,
clear, and empty, as well as a number of denominal verbs (i.e., verbs derived from
nouns, including mop and shovel. Which subclass are the deadjectival verbs found
in, the means/manner subclass or the result subclass? Which subclass are the
denominal verbs found in? Why are these verbs found in the particular subclasses
you just identified?
(c) In “The Grammar of Hitting and Breaking”, Fillmore argues that the objects of
break and hit do not bear the same semantic roles. Do you think the sidewalk
should be assigned the same semantic role in both Robin cleared the sidewalk and
Robin shoveled the sidewalk ? As part of your answer, address whether either or
both uses of the sidewalk should be assigned the Patient role. Give evidence that
supports your conclusion.

Part V: Properties and the mass/count distinction

Answer the following question. [20 points]

Question 9: Adjectives modifying mass nouns

(a) The examples in (14) show that only some of the semantic classes of adjectives can
modify mass nouns. Which classes of adjectives are and aren’t found before mass
nouns, according to the examples below? Why might these differences in types of
modifiers be found between mass and count nouns?

(14) a. Chinese jade, clear water, Danish butter, fresh cream, good coffee, gray
stone, Greek honey, hot cider, Indian tea, Italian leather, old glass, pink
granite, purple ink, red clay, soft butter, Spanish oil, tasty mustard,
white chocolate
b. *big marble, *hexagonal copper, *large mud, *long silk, *narrow sand,
*oval gold, *rectangular chocolate, *round jam, *tiny ruby

(b) Some English mass nouns can be found with the kinds of adjectives that the mass
nouns in part (a) cannot co-occur with: big jewelry, miniscule cutlery, triangular
furniture, large clothing. Why are these nouns different from those in (14)?