The Loss of V2 Clause Structure in the History of English:A Review of the Literature
Heather Burnett and Elizabeth Ferch
Queen’s University
August 24, 2005
© Summer 2005
Cette recherche a bénéficié de l’appui du projet Modéliser le changement : les voiesdu français (GTRC/CRSH), dirigé par France Martineau. La recherche a étésupervisée par Mireille Tremblay, Queen’s University.
Table of Contents_____________________________________________________________
Section 1: Review of the Literature
0. Introduction…………………………………………………………………... 31. Theories of Language Change……………………………………………….. 32. Synchronic Analyses of Old English……………………………………….. 143. Synchronic Analyses of Early Middle English……………………………... 274. Late Middle English and the Loss of V2…………………………………… 305. Conclusion………………………………………………………………….. 33
Section 2: Article Summaries…………………………………………………………... 35
Section 3: Bibliography of V2 in English……………………………………………... 114
Section 1: Review of the Literature
0. Introduction
Historical linguistics and the study of language change was once the major area of
study for linguists. With the advent of generative grammar, it has taken on a new role.
The study of the way in which language changes, and the way in which it doesn’t, offers
a unique window into the constitution of the language faculty. In this paper, we will
present a comprehensive and critical review of the literature surrounding a major change
in the history of the English language: the loss of verb second word order from Old
English to Middle English. We will begin by describing some of the main theories of
language change which have been used to study the loss of V2. We will then examine
the debate on the synchronic description of word order in Old English. Finally, we will
review the various analyses of the synchronic state of Middle English, as well as the
proposals for the causes and time course of the loss of V2. The goal of this paper is to
make accessible the vast amount of literature available on the subject, to fuel further
research into the history of English and diachronic syntax in general.
1. Theories of Language Change
In this first section, we will present a brief description of some of the major
theories of language change that the discussion on the history of V2 has been couched in.
Although they cannot be entirely separated, we feel that it is very important to make a
distinction between the synchronic descriptions of various stages of a language, and the
way in which the transition between those stages is expressed. This is because a certain
theory of change can hold valuable insights, even if the particular synchronic analysis
used to exemplify it may not be 100% correct.
Roberts (1993) distinguishes three components in his theory of language change.
A step, defined as “the appearance of a new construction, or a significant change in the
frequency of a construction” (Roberts 1993: 158), is a diachronic relation between E-
languages; a parametric shift is a diachronic relation between I-languages; and a
Diachronic Reanalysis (DR), the reanalysis of a given structure, is a relation between the
E-language of one generation and the I-language of the next. Diachronic Reanalyses are
caused by the Least Effort Strategy, which prompts learners to choose the grammatical
representation containing the shortest possible set of chains, and they often provoke
parametric shifts by eliminating the evidence for a given parameter setting; Roberts
suggests that they are always minimal, that is, that each reanalysis reduces the structure
as little as possible. Although steps and DRs can reduce the frequency of constructions,
only parametric shifts can make them ungrammatical. However, as Lightfoot (1997)
points out, the distinction between a parametric shift and a Diachronic Reanalysis is not
entirely clear; Roberts himself allows that “[a]ll DRs may turn out to be instances of
Parametric Change” (1993: 159). As well, this theory does not explain why the LES
should provoke a reanalysis in one generation of learners and not another.
Lightfoot (1991, 1995, 1997) proposes a theory of language change based on
sudden reanalysis. He postulates that parameter settings are acquired based on specific
triggering experiences in the linguistic environment. Based on data from Dutch and
German regarding the acquisition of OV word order (Klein (1974); Clahsen & Smolka
(1986), cited in Lightfoot (1991)), and data from its loss in Old English (Bean (1983) and
others), Lightfoot argues for “degree-0 learnability,” in other words, the idea that
triggering experiences must be present in simple matrix clauses. Triggers must also be
expressed robustly, that is, in a sufficient proportion of sentences, in order to cause the
parameter to be set. Lightfoot believes that the linguistic environment is always
changing, due to grammar-external factors; in cases of language change, a particular
construction increases or decreases in frequency until the triggering experience has
changed sufficiently for the parameter it expresses to be set a different way. Thus,
Lightfoot’s theory involves a gradual drift in usage followed by a sudden change in
grammar. However, Kroch (1999) takes issue with this notion of “drift”, saying that
empirical evidence for its existence is “at the least, uncertain” (Kroch, 1999:6). He claims
that the problem lies in the fact that so-called “drifting” has no obvious motivation;
therefore, it is unclear why the frequency of a certain form undergoing change should be
incorrectly learnt in one case, but correctly learnt in practically all the other cases. He
cites cases such as the stability of the frequency of the placement of adverbs in the history
of English as an example where drift would have most likely occurred, but did not
(Kroch, 1999:6-8). According to him, if the drift model was correct, the loss of V to I
movement in Middle English would have caused an increase in the frequency of
preverbal adverbs. However, the frequency of preverbal adverbial placement versus
postverbal placement remained the same throughout time, thus rendering the motivation
for the change in frequencies that bring about a reanalysis extremely unclear.
In order to avoid postulating drifts in usage frequencies that occur prior or
independently of grammar change, Kroch (1989; 1999, 2003 a.o.) proposes the
“imperfect transmission” (IT) model of language change. In this model, instead of the
grammatical change occurring at the end of a long period of drift, it is the first stage in
the process of change. According to the “IT”, change occurs when, for a variety of
reasons, a child “mislearns” the grammar of its parents. They create a grammar that is
different from the one they are exposed to, and it is this difference in grammar that
creates a difference in the patterns of usage.
Basing the model in the principles and parameters (P&P) framework (for an
overview, see Chomsky, 1995: chapter 1), Kroch (2003), following Clahsen (1991) and
Penner (1992; both cited in Kroch, 2003), assumes that language learners set linguistic
parameters in an irreversible way. Therefore, when the learner first encounters some
input that makes reference to a certain parameter, she will set it accordingly. If all the
following input that she is exposed to is consistent with this parameter setting, the learner
will simply assume that this is the target grammar and nothing further will happen.
However, if, later on, she comes into contact with data which contradicts the first
parameter setting, seeing as she cannot simply re-set the parameter, nor can she put the
parameter on both settings, the learner creates a separate grammar which has the
parameter in question set in the opposite way. The learner is now “bilingual” or
“bidialectal”, in the sense that she has two (or more) separate grammars in her head and
can use either one of them. The two grammars do not, however, coexist completely
harmoniously; they are in what Kroch (1989 and later work) and Pintzuk (1991 and later
work) call “grammar competition”. In a first language acquisition sense, this term
corresponds to a situation where a child mistakenly acquires a parameter setting that is
clearly different from the majority of input sentences she is exposed to. According to
Kroch (2003), if the learner is young enough and the primary linguistic data conclusively
support the second, newer grammar, then she will begin using the second grammar at the
expense of the first. After a period of variation/competition, the second grammar will
“win out” over the first, i.e. it will become the dominant one all but eradicating the first.
Kroch (2003) claims that a concrete example of this type of grammar competition is
found in the study of Labov & Labov (1976). In this study, the Labovs recorded
thousands of their four-year old daughter’s wh-questions and note that the acquisition of
questions formed with “why” undergoes a radically different pattern with respect to the
presence of inversion than the other Wh-words:
(1)
Acquisition of Wh-questions by Labov daughter (cited in Kroch (2003:his
figure1)
Kroch interprets these results in the following way: until roughly December 1976, the
Labovs’ daughter had inaccurately set the “inversion” parameter as – in the case of why-
questions, but as + in the case of the other wh-questions. Having been exposed to
overwhelming evidence of the existence of a grammar with the +inversion setting, she
then creates one which enters into competition with the existing –inversion grammar. The
+inversion grammar ends up replacing the original grammar by June, 1976.
Kroch (1989 and subsequent work) proposes a model of language change which
essentially mirrors the process of grammar competition in acquisition, but on a larger
scale. According to this model, language change occurs when the primary linguistic data
available to learners consistently requires the creation of two grammars, the competition
between which is not eradicated within the critical period of acquisition. Speakers can,
therefore, live their entire lives with multiple grammars, and they make choices as to
frequencies at which they use one grammar over the other that are “probabilistically
influenced by features of context and situation” (Kroch, 1989:3).
Following authors such as Weinreich, Labov and Herzog (1968, cited in Kroch,
1989) and Osgood and Sebeok (1954 cited in Kroch, 1989), Kroch assumes that the time
course of language change follows an ‘S’-shaped curve, with “new forms replacing
established ones only slowly in the beginning of a change, then accelerating their
replacement in the middle stages of a change and finally, as the old forms become rare,
slowing their advance once again” (Kroch, 1989:3). Kroch (1982, 1989) goes even
further in his characterization of the shape that language change takes by proposing a
specific mathematical function, the logistic1, as the underlying form of a syntactic
change. Although, as Kroch (1989:4) notes, it is impossible to verify exactly which
function is the proper characterization of the time course of language change, other
1 The equation of the logistic curve is where p is the fraction of theadvancing form, t is the timevariable, and s and k areconstants.
models that have been used, such as the cumulative function of the normal distribution
(Aldrich and Nelson, 1984, cited in Kroch, 1989), “generally differ so little from the
logistic that they can provide no improvement in fit to empirical data.” Also, the logistic
has been widely used in other fields to model population behaviour, such in epidemiology
and population biology (ex. Spiess, 1989, cited in Kroch (1989); therefore, the logistic
seems to be a plausible model to apply to grammar competition and replacement.
Further support for parameter-based grammar competition comes from the rate
that the time course of language change takes. If one assumes that the grammars in
competition differ in the setting of an abstract parameter, one would expect to see the
result of a parametrical difference to take the form not of a single difference in form, but
as a series of correlated changes that instantiate the abstract change. Therefore, we would
expect all the elements affected by the parametrical change to be change at the same rate.
This is known in Kroch (1989 and subsequent work) as the “constant rate hypothesis”.
More specifically, it is the claim that, although the rate of use of different grammatical
options in competition will most likely differ across contexts at each period in time, the
rate of change will be the same across contexts (Kroch, 1989:6). The realization of this
hypothesis can be illustrated through the case of the loss of V2 in the history of French.2
As shown in (2), the losses of pro-dropped subjects, full NP subject inversion, and
pronoun-subject inversion proceed at the same rate, as would be expected if they were the
result of a change in a certain “V2 parameter”.
2 We bring up this case, as Kroch does, simply for explanatory purposes. At this point we do not support ordeny this claim about the actual nature of V2 in French; however, for a deeper discussion about the issuesinvolved in the history of French V2 see Burnett & Ferch (in prep).
(2)
(Kroch, 1989: 12, his Figure 3)
Another approach that strongly links language change with acquisition is that of
Yang (2000). Yang compares the tension that exists between the discrete and symbolic
nature of grammatical competence and the variable, heterogeneous nature of performance
with the tension that exists between the discrete basis of Mendelien genetic and the
continuous distribution of genotypes in populations. He views language acquisition as
being similar to evolution: it is as a “variational process in which the distribution of
grammars changes as an adaptive response to the linguistic evidence in the environment”
(Yang, 2000:234). As in Kroch’s model, this “variational process” is realized through
multiple grammar competition and replacement. In this model, within the finite collection
of grammars that make up UG, each grammar Gi is associated with a weight pi, which
denotes the probability with which the learner will access the grammar. Acquisition
works as follows: when an input sentence is presented, a grammar is selected. The
probability that certain grammar is selection is determined by its weight. The grammar is
then used to analyze the sentence. If the sentence can be successfully parsed using the
selected grammar, it is rewarded: i.e. its weight is increased, and the weights of all the
other grammars are indirectly punished: i.e. their weights are decreased. If the selected
grammar fails to parse the input, it is punished and the other grammars are rewarded. This
process continues until a single grammar has a weight that ensures that it will be
exclusively selected.
If we assume, as is common, that UG remains constant from generation to
generation, then language change must be due to a situation where two generations are
exposed to different linguistic evidence, which results in different knowledge of
language. According to Yang, the “different linguistic evidence” can arise in a variety of
ways such as through the migration of foreign speakers who introduce novel forms into
the language, through linguistic innovation, or through more general social and cultural
factors (237). These factors will create a situation where the input evidence will seem to
reflect two conflicting grammars. The proportion of expressions that belong to Grammar
1 (G1) that are incompatible with Grammar 2 (G2) is called the advantage of G1, and the
proportion of expressions in the environment that belong to G2 that are incompatible with
G1 is called the advantage of G2. Therefore, according to Yang’s theory of acquisition,
Grammar 2 will overtake Grammar 1if the advantage of grammar G2 is higher than G1.
The change takes the shape of the logistic, as shown in (3)
(3)
(Yang 2000: 240)
Due to the repeated “punishment” of unsuccessful grammars that it proposes, a corollary
of Yang’s model is that, once a grammar is on the rise, it is unstoppable (239).
Another theory which takes an evolutionary look at language acquisition and
change is that of Clark and Roberts (1993), who characterize language acquisition as
natural selection among hypotheses about the target grammar. In their theory, hypotheses
are made up of sets of parameters and encoded as strings of binary digits. Each
hypothesis translates into a parsing device; these parsing devices are tested against each
input datum, and their fitness is judged based on how they interpret the datum3. The most
3 C&R provide a fitness metric for mathematically calculating the relative fitness of the grammars. Thismetric is shown in (i):
(i) n n n(_j = 1 vj + b_ j = 1 sj + c_ j = 1 ej) – (vi + bsi + cei)n n n(n – 1)( _j = 1 vj + b_ j = 1 sj + c_ j = 1 ej)
n
n n
important factor is the well-formedness of each parse, specifically the number of basic
grammatical principles which are violated. Other factors are also considered, including
the elegance of the parse, measured as the number of nodes contained in it, and the
number of superset parameters; these factors are important when distinguishing among
equally grammatical interpretations. A superset parameter is one which, when set a given
way, produces a set of sentences which includes all sentences that could be generated
with the parameter set the other way. If the learner were to incorrectly postulate the
superset setting, there would be no evidence available to disprove this hypothesis, and the
language acquired would be a superset of the target language; in order for this not to
happen, such hypotheses must be selected against, as per the Subset Condition of
Berwick (1985). When the relative fitness of the hypotheses is determined for a given
datum, the most fit hypotheses are selected for special “mating operations”: a crossover
mechanism, by which a new hypothesis is formed from parts of two old ones, and a
mutation operator, which forms new hypotheses by randomly changing parameter
settings. Over time, parameter settings which are frequently expressed in the input will
be selected for, while those which are unable to interpret the input will be selected
against; thus, the population of hypotheses will tend towards the optimal settings for the
target language. This process does not, however, guarantee that the learners’ grammars
will be identical to those of their parents. If a parameter is unambiguously expressed in
the language, it is stable, but if the evidence becomes ambiguous between various
In this equation, n represents the number of parsing devices; _j = 1 vj represents the total number ofgrammatical violations of all parsing devices; vi represents the number of violations by the parsing devicewhose fitness is being determined; _j = 1 sj represents the total number of superset settings in the populationof hypotheses; si represents in number of superset settings in the grammar being evaluated; _j = 1 ejrepresents the total number of nodes generated by all parsers; ei represents the number of nodes (theelegance) of the parser being evaluated; and b and c are constants, greater than 0 and smaller than 1, used toweight the elegance and superset factors so that they have less impact than does failure to parse.
settings, the language will be unstable and therefore prone to change. Instability can arise
for various reasons: phonological changes may make syntactic parameter settings
difficult to interpret; a change in one parameter may destabilize another one; or the
language may become “shifted,” with learners setting parameters in such a way that they
interact to produce a superset language. Once a language becomes unstable, according to
Clark & Roberts, learners turn to UG-internal factors, specifically elegance and the
Subset Condition, to choose between hypotheses.
This model is similar toYang (2000) in that, as in his model, the outcome of
language acquisition is determined by the compatibilities of grammars with the linguistic
input data in a “Darwinian selectionist manner” (Yang, 2000:240), the differences
between the two models being primarily that Clark and Roberts (1993) do not accept, like
Kroch and Yang, the existence of multiple grammars within the head of a single speaker.
These evolutionary theories are impressive, but as Kroch (1999) notes, theories such as
that of Clark and Roberts “are useful hypotheses, no doubt, but unless they can be further
specified to make empirically testable predictions, they will remain speculative”
(1999:35).
2. Synchronic Analyses of Old English
The proper characterization of Old English syntax has been a matter of great
debate. Due to the large amount of complex variation in its surface word order, a number
of scholars have concluded that OE was a free word order language (Kroch & Taylor,
1997). For those who noticed the V2 pattern in OE, it was still uncertain whether or not
this was a robust syntactic rule, or just a narrative tendency (see Bean, 1983). However,
in more recent years, studies such as those done by van Kemenade (1987) and Pintzuk
(1991) have shown that OE word order in general and, more specifically, the V2 pattern
is, in fact, highly rule governed. However, the exact nature of the V2 rule remains open to
discussion. In this section we will first give a brief overview of the V2 pattern in OE, and
then we will present various analyses which have tried to account for the distribution of
verb-second order in both main and subordinate clauses.
If, following van Kemenade (1987), Pintzuk (1991;1993), Kiparsky (1994), Kroch
& Taylor (1997), we take the dominant underlying order of late OE to be SOV, then the
most common type of V2 sentence would be the one where the finite verb moves to
become the second element in the clause, and the NP-subject is the topic, yielding surface
SVO order, as in (4):
(4) Godes _eowas sceoloni unscæ__ignysse healdan tiGod’s servants should harmlessness preserve‘God’s servants should preserve harmlessness…”
(AELS 25.846, cited in Pintzuk, 1993: 6)
However, the subject is not the only thing that can appear before the verb. It is also very
common to have the “topic” position filled by a non-subject: a direct object, a
prepositional phrase or an adverb. In this case, the subject appears after the verb, as
shown in (5):
(5) a. & of heom twam is eall manncynn cumen (WHom 6.52) and of them two is all mankind comeb. pæt hus hæfdon Romane to dæm anum tacne geworht (Or 59.3) that building had Romans with the one feature constructedc. pær wearp se cyning Bagsecg ofslægen (Anglo-Saxon Chron., Parker, 871) there was the king Bagsecg slain
(all cited from Kroch & Taylor (1997:302), their
(5))
Another pattern that we see with non-subject topics is “verb third” word order. This arises
in cases where the subject of the sentence is a pronoun:
(6) Ælc yfel he mæg don (WHom, 4.62) each evil he can do
(cited in Kroch & Taylor, 2000, their (7a))
Although V3 word order was originally taken to indicate the inconsistent nature of the V2
constraint in OE, Kemenade (1987) and Pintzuk (1991) have shown that it reflects not a
variable verb movement rule, but particular characteristics of the pronouns themselves (to
be discussed below). Evidence to suggest that V3 is a result of the “clitic-like” nature of
the pronoun and not of the V2 constraint comes from the fact that object pronouns can
appear in the same preverbal position as their subject counterparts.
(7) & seofon ærendracan he him hæfde to asend (ASC, Parker, 905)and seven messengers he him had to sent‘and he had sent seven messengers to him’
(cited in Kroch, Taylor & Ringe, 2000, their (8b))
V3 order can also occur when some adverbs are topicalized, as in (8):
(8) Her Oswald se eadiga arcebisceop forlet pis lif. (ASC, Laud, 992)in-this-year Oswald the blessed archbishop forsook this life
(cited in Kroch, Taylor & Ringe, 2000, their (10c))
V2 order also appears in a small number of exceptional constructions such as non-
subject wh-questions, sentences introduced by the elements _a and _onne, and negation-
initial sentences, examples of which are found in (9):
(9) a. hwi sceole we opres mannes niman? (AELS 24.188) why should we another man's takeb. pa ge-mette he sceadan (AELS 31.151) then met he robbersc. ne mihton hi nænigne fultum æt him begitan (Bede 48.9–10) not could they not-any help from him get
(cited in Kroch, Taylor & Ringe, 2000, their (9a-c))
These constructions are customarily analyzed as involving verb movement to C (Pintzuk
(1991, 1993), Kroch & Taylor (1997), Kemenade (1987, 1997), Fischer et al (2000)).
As for subordinate clauses, the data are much less clear. According to Pintzuk
(1991), Old English exhibits V2 order in a variety of subordinate clauses. In her view,
this makes OE different from other V2 languages such as German and Dutch which allow
verb second only in main clauses4. This significant difference leads Pintzuk to postulate
that OE is an IPV2 language (like Icelandic and Yiddish) rather than a CPV2 language,
such as German and the Mainland Scandinavian languages. It is generally assumed that in
the latter languages the verb raises to the complementizer position, and the topic moves to
SPEC, CP. However, when the C position is filled by a complementizer, as is the case in
subordinate clauses, verb movement to C is blocked, yielding a non-V2 word order. In
an IPV2 language, the verb raises to the inflectional head I, and the topic moves to SPEC
IP. In subordinates, V2 order is still possible since the I head is not affected by the
presence of a Comp in C. In order to explain the V3 effect that arises when the subject is
a pronoun, Pintzuk proposes that pronouns in OE are IP (Wackernagel) clitics, which
means that they attach at the border between CP and IP. She also postulates the existence
of a phonological rule which moves them to the right of the first phrasal constituent,
yielding the order found in (6).
Pintzuk’s analysis is reviewed in Kroch & Taylor (1997) (K&T), and, although
they largely support her claims, these authors raise a number of concerns with her
proposal. Primarily, they take issue with the phonological rule that is needed to derive the
surface XP Pro V order. As they say, “the special clitic movement rule needed by Pintzuk
4 This does not take into account special cases which can be analysed as instances of “CP recursion” asdiscussed in Iatridou & Kroch (1992).
to account for the placement of pronouns between topic and verb in V2 clauses has no
counterpart elsewhere among the Germanic languages and does not have clear theoretical
justification” (Kroch & Taylor, 1997: 305). We would also like to point out that the
“phonological” nature of this postposing rule would predict that that the final word of the
first constituent and the subject clitic would form a phonological word. For example, in
(6) (repeated here as (10)), Pintzuk’s rule would have “he” forming a phonological word
with “yfel”, which seems to us to be counterintuitive.
(10) Ælc yfel he mæg don (WHom, 4.62) each evil he can do
(cited in Kroch & Taylor, 2000, their (7a))
In order to remedy these inconsistencies, K&T propose that, in Old English main clauses,
the topic moves first to SPEC, IP but goes on to land in SPEC, CP. As such, there is now
no need to propose a phonological rule that moves the clitic from the Wackernagel
position at the CP/IP border. K& T do, however, support Pintzuk’s claim that the topic
moves only to SPEC, IP in subordinate clauses.
The second difficulty that K&T have with Pintzuk (1991) is that it fails to account
for a striking generalization made from data presented in van Kemenade (1997): rather
than being relatively productive (as Pintzuk’s theory predicts), V2 word order with a non-
subject topic in subordinate clauses is possible only in cases where “the nominative
subject is absent or is licensed to appear in a position other than SPEC, IP, as in passive
sentences or in sentences with experiencer dative “subjects” ” (K&T: 307). These two
kinds of V2 subordinates are shown in (11):
(11) a. passive:_onne ælce dæge beo_ manega acennede _urh hys mihte on woruldwhen each day are many(NOM) given-birth through his power on world
(AEHP.VI.120, cited in K&T:307)
b. experiencer dative “subjects”:gif us ne lyst _æra ærrena yfela _e we ær worhtonif us(DAT) not pleases the earlier evils (GEN) that we earlier wrought
(CP.445.29, cited in van Kemenade (1997: 335))
Thus, according to K&T, SPEC, IP is available for topics only when the subject does not
have to be there. Both Pintzuk (1991)’s analysis and K&T’s refinement are immediately
confronted with the following questions: why is it that the subject is not required to move
to SPEC, IP in order to check case and EPP features in main clauses? Conversely, if the
subject is not required to move to SPEC IP in main clauses, why is it that it must raise in
subordinates?
K&T’s solution to these problems is the following: If one takes the view that the
essence of the V2 constraint is the establishment of a SPEC/head relationship between the
topic and the verb, and if the verb only moves to I, it is then imperative that the topic
occupy the specifier position of IP.5 Since the topic or its trace is in SPEC, IP, the subject
cannot move there. In order to satisfy the Principle of Full Interpretation (Chomsky
(1986; 1993), it must check the finite agreement features in another way. According to
K&T, this is accomplished through the incorporation of a null expletive which is
coindexed with the subject into the feature complex of I. This expletive checks the active
inflectional and EPP features in I, making it unnecessary for the subject to move into
SPEC, IP. In subordinate clauses, the movement of the subject to SPEC, IP is not driven
by inflectional or EPP considerations; it is, in fact, due to the topicalization that is
necessary to satisfy the V2 constraint. Although this scenario addresses the theoretical
problem of the licensing of the subject inside the VP, K&T themselves note that “since
the licensing of topics in main and subordinate clauses is identical, we are now without a 5 K&T note that It is also possible, assuming a split IP comprised of AgrS and T, that the verb moves toAgrS. (K&T, 1997:319).
simple syntactic explanation for the greatly reduced range of topicalizations in
subordinate as opposed to matrix clauses” (K&T: 309). However, K&T suggest that this
restriction may in fact not be syntactic but pragmatic: they claim that due to information
structure considerations, the subject will be topicalized as much as possible in
subordinate clauses, the only exceptions to this being the cases, such as in (11), where
there seems to be no subject. Although it is impossible to verify this hypothesis for OE,
evidence from Icelandic and Yiddish (Rognvaldsson & Santorini, p.cs to K&T) suggests
that this generalization holds in living IPV2 languages.
Another analysis of the Old English syntax is that of van Kemenade (1987;1997)
who proposes that OE is, in fact, a CPV2 rather than IPV2 language. She claims that, in
main clauses, the topic moves to SPEC,CP and the verb moves to the complementizer
position. Pronouns cliticize on the left of the finite verb, as in (12a), but the cliticization is
blocked if the first constituent is an operator (i.e. contexts where the verb independently
raises to C, such as in wh-questions) (12b):
(12) a.Be _æm we magon swi_e swutule oncnawan _æt… (CP.181.16) by that we may very clearly perceive thatb. for hwam noldest _u _e sylfe me gecy_an (AESLXXXIII.307) for what not-wanted you yourself me make known ‘wherefore would you not make yourself known to me’
(cited in van Kemenade (1997:333)
By saying that the topic moves to SPEC, CP in main clauses, van Kemenade, like Kroch
& Taylor (1997), eliminates the need to postulate the existence of a special phonological
clitic postposition rule of the type argued for in Pintzuk (1991). However, unlike Pintzuk
(1991) and K&T, van Kemenade does not suggest that the topic is ever in SPEC, IP. This
placement of the topic predicts that V2 word order would be impossible in subordinate
clauses. Van Kemenade (1997) argues that this is, in fact the case, and she shows that
apparent V2 order in subordinates only occurs in constructions which have no external
argument. This distribution was taken by K&T to reflect a pragmatic constraint on
embedded topicalization rather than a syntactic one, and they analyze cases such as the
ones in (11), repeated here as (13), as real instances of V2 in subordinates.
(13) a. passive:_onne ælce dæge beo_ manega acennede _urh hys mihte on woruldwhen each day are many(NOM) given-birth through his power on world
(AEHP.VI.120, cited in K&T:307)b. experiencer dative “subjects”:gif us ne lyst _æra ærrena yfela _e we ær worhtonif us(DAT) not pleases the earlier evils (GEN) that we earlier wrought
(CP.445.29, cited in van Kemenade (1997: 335))
However, van Kemenade (1997) argues that these cases do not constitute valid examples
of V2 in subordinates as it is impossible to tell that what K&T are calling the “topic” (i.e.
“each day” in (13a) and “us” in (13b)) is actually in SPEC, IP. For example, it is not clear
that “each day” is necessarily in SPEC, IP, since, as shown in (14), there is a position
between SPEC,IP and I available for temporal adverbials.
(14)
Kroch (2005:L6:3)In order to properly conclude that the V2 constraint is active in subordinate clauses, one
would have to come across a case where the external argument was postverbal with a
non-subject XP occupying the position of SPEC, IP. Van Kemenade (1997) says that,
since cases such as in (13) do not have an external argument, they cannot be used as
evidence for the presence of the V2 constraint in subordinates. We also think that it is
possible that the constituents which K&T classify as “non-subject topics” in sentences
such as (13b) are actually subjects. K&T conclude that the DP “us” is not generated in
SPEC, vP based on its dative case marking. However, it has been shown for Icelandic
(Harley, 1995) that there are a certain group of verbs that lexically assign dative case to
their subject and nominative case to their object, such as the verb “liki” in (15):
(15) Calvini liki verki_Calvin-D like the job-N“Calvin likes the job”
(Harley, 1995, her (1))
Despite their “quirky” case marking, these DPs display properties of subjecthood (for a
complete discussion, see Harley (1995); Bobaljik (2005)). As shown in (16), it seems
likely that there is a similar group of verbs in OE, of which “licode” is a member:
(16) ac Gode ne licode na heora geleafleastbut God(DAT) not pleased not their faithlessness(NOM)‘but their faithlessness did not please God’
(AEHP.XX.71, cited in van Kemendade (1997:334))
In these sentences, SPEC,IP would be occupied by the subject, and thus examples such as
(13b) give us no information at all as to the presence or absence of V2 in subordinates.
Given the ambiguous evidence, it is difficult to uphold an IPV2 analysis of Old English;
however, until these constructions are examined in more depth it cannot be completely
ruled out.
Van Kemenade’s (1987;1997) analysis is slightly revised in Fischer et al. (2000).
This revision concerns the treatment of pronouns in OE. They propose that OE pronouns
are Germanic-style weak pronouns and not clitics, as claimed by Pintzuk (1991;1993),
Kroch & Taylor (1997), Van Kemenade (1987; 1997), and Cardinaletti & Roberts (1991).
Following Koopman (1997), they claim that these pronouns are housed in the specifier of
a functional projection (FP) that is below C, but above IP and the high negation phrase. In
main clauses, the topic moves to SPEC, CP and the verb moves to the head of FP to join
the pronouns.
(17)
Based on Fischer et al. (2000:126)
Treating the pronouns as independent syntactic objects rather than as clitics allows
Fischer et al. (2000) to account for the fact that pronouns appear after the verb in contexts
such as Wh-questions. In these cases, the verb raises past the pronouns to C yielding the
WH V Pro order. There is thus no need for an extra rule concerning the direction of
cliticization, such as that in van Kemenade (1987;1997). Haeberli (2002) has a very
similar proposal to the one made in Fischer et al. (2000). The structure is the same; the
only difference being that Haeberli calls FP AgrP. He proposes that subjects below AgrP
are licensed by an empty expletive in SPEC, AgrP.
A final, somewhat different, analysis comes from Cardinaletti and Roberts (1991)
(C&R). They argue for a more articulated clause structure than is usually assumed, which
includes two subject agreement projections.
(18)
Based on (C&R, their (0))
The lower agreement projection holds the agreement morphology of the verb, and the
upper one holds subject clitics. In main clauses, the verb joins the clitic in Agr1;
however, in subordinate clauses, it only moves as far as Agr2.
It is difficult to evaluate this analysis in comparison with the others as it is an
analysis more of verb movement than it is of the V2 phenomenon. C&R make no explicit
mention of the position of topics or what the relationship between the topic and the verb
is. Also, the expanded clause structure allows for an easy derivation of a lot more
possible word orders than do the analyses of Pintzuk (1991), Kroch & Taylor (1997),
Kemenade (1987), and Fischer et al. (2000). However, this account may not be the most
interesting one to pursue, as the existence of agreement projections has been argued
against by Chomsky (1995:349-355), who notes that it is motivated solely by theory-
internal reasons. In the interests of empirical falsifiability, it is thus desirable to postulate
the existence only of projections “with intrinsic properties that are manifested at the
interface levels” (Chomsky, 1995:355), not ones, such as the many Agr projections in
C&R, which have no semantic or phonological consequences.
The various approaches to the syntax of Old English examined here are
summarized in table (1):
Table (1): Analyses of V2 in Old EnglishPintzuk (1991;1993)
Kroch & Taylor (1997) Van Kemenade(1987; 1997)
Fischer, vanKemenade, Koopman& van der Wurff(2000)
Cardinaletti &Roberts (1991)
Mainclauses
Topic SPEC, IPVerb I
TopicSPEC, CP (passingthrough SPEC,IP)Verb I/AgrS
Topic SPEC, CPVerb C
Topic SPEC, CPVerb F
VerbAgr1
Subordinateclauses
Topic SPEC, IPVerbI
Topic SPEC IPVerb I/AgrS
Not V2 Not V2 VerbAgr2
Status ofpronouns
IP clitics + aphonological rulepostposing themto the right of thefirst constituent
IP clitics IP clitics Germanic-style weakpronouns housed inSPEC, FP, not clitics
Clitics in Agr1
In conclusion, the analysis given by Fischer et al. (2000) and Haeberli (2002)
seems to be best able to account for the various word order patterns in OE in a straight-
forward manner; however, we are concerned that, through the addition of the extra
functional projection needed to account for the distribution of pronouns, they lose what is
often considered the motivation for V2: the establishment of a local (Spec/head)
relationship between the topic and the verb (Kroch & Taylor, 1997). It is therefore clear
that, at the moment, there is no completely satisfactory analysis of the V2 phenomenon in
Old English, and the question is still open for discussion.
3. Synchronic Analyses of Early Middle English
V2 word order continues to be dominant well into the Early Middle English
period. However, in this time period, we can see the development of a dialectal difference
in the V2 constraint with respect to the behaviour of pronouns. The Southern dialects
exhibit the same pattern as in Old English: subject-inversion occurs with full NPs and not
with pronouns. The Northern dialects, however, start to show inversion with all types of
subjects, as shown in (19).
(19) O_ir labur sal _ai doother labour shall they do‘they must do other labour’ (Benet.33.20 cited in Van Kemenade (1997))
For Van Kemenade (1987, 1997), verb movement in Old and Early Middle
English is basically the same. According to her, Old English is a CPV2 language, a
pattern which continues into the 14th century. Van Kemenade (1997) attributes the
difference in the position of pronouns between the northern and southern dialects to a loss
of a clitic position in the North which remains available in the South.
The Loss of V2 in the History of English 28
However, Kroch & Taylor (1997) take issue with this simplistic approach, saying
that it does not account for another change occurring at the same time in the Northern
dialects: the increase in inversion of full NPs with adverbs and adjunct PPs. As shown in
(20), there is a higher rate of subject-inversion in the Northern dialects than in the South
and Midlands.
(20a)
Table (2): subject-verb inversion in the Northern Prose Rule of St. Benet(from Kroch & Taylor (1997:313)
(b)
Table (3): subject-verb inversion in seven early Midlands texts(from Kroch & Taylor (1997:311)
This dialectal difference cannot be explained by a simple difference in the cliticization of
pronouns. Also, there is reason to believe that pronouns in the “Rule of St. Benet”, a
Northern text, are still clitics. For example, stylistic fronting can occur with pronouns, as
in (21), but not with full NPs.
The Loss of V2 in the History of English 29
(21) Bot yef it sua bi-tide, _at any falle in mis-trouz; _an sal scho pray gerne to godbut if it so betide, that any fall into mistruth; then shall she pray earnestly to God
(Benet 19.30, cited in Kroch & Taylor 1997: 314)
In order to account for these differences, Kroch & Taylor propose that the IPV2 system
which they postulated for Old English is replaced by a CPV2 system in the Northern
dialect, while the Southern dialect remains IPV2. This means that in the North, the verb
moves to C, and the topic moves directly to SPEC,CP, not leaving a trace in SPEC,IP as
it did in Old English. In this case, the verb in C always precedes pronouns cliticized at
the CP/IP boundary; the increased rate of inversion with adverbs and adjunct PPs is
explained by the fact that adjunction to CP is less productive than adjunction to IP, so
that in the Northern CPV2 dialect, it is more likely that initial adverbs or PPs are topics
rather than preverbal adjuncts, and therefore that they trigger inversion.
The fact that a change from IPV2 to CPV2 can account for the differences
discussed above seems to provide evidence for an IPV2 characterization of Old English;
however, the increased rate of subject inversion with adverbs and PP adjuncts is not
necessarily due to such a change, as it might simply reflect an increased preference for
topicalizing such constituents in the Northern dialect. Thus, a CPV2 analysis could still
be upheld. However, the asymmetry between pronouns and full NPs when it comes to
stylistic fronting remains a problem for van Kemenade’s (1997) decliticization analysis.
This asymmetry can be accounted for by the analysis of Fischer et al (2000).
They adopt K&T’s idea of a dialectal difference between North and South, with a CPV2
dialect in the North, but they characterize the Southern dialect as having verb movement
to F, as in their analysis of Old English, rather than to I.
The analyses we have discussed are summarized in Table (4):
The Loss of V2 in the History of English 30
Van Kemenade(1997)
Kroch & Taylor(1997)
Fischer et al(2000)
Main clauses: North Topic SPEC, CPVerb C
Topic SPEC,CPVerb C
Topic SPEC,CPVerb C
South Topic SPEC,CPVerb C
Topic SPEC,CPVerb AgrS
Topic SPEC,CPVerb F
Subordinates: North Not V2 (Topic SPEC,CP)6
Verb TNot V2
South Not V2 Topic SPEC, IPVerb T
Not V2
Pronouns: North Not clitics Clitics Weak pronouns South clitics Clitics Weak pronounsTable (4): Synchronic analyses of Early Middle English
4. Late Middle English and the Loss of V2
The number of unambiguously verb second sentences gradually declines starting
from the middle of the 14th century (Kroch, Taylor & Ringe, 2000), although it is not
completely eradicated until the 17th century (Fisher et al. 2000), as shown in (22):
(22)
Frequency of subject-verb inversion (V2) in the Penn-Helsinki Parsed Corpus of MiddleEnglish and Early Modern English (from Johnson & Whitton (2002), cited in Kroch(2003).
6 In Middle English, it is possible to have topicalization across a lexical complementizer in ungovernedsubordinate clauses, as in (i).
(i) I sal yu lere _e dute of god, his wille _at Ζe may doI shall you teach the duty of God, his will that ye may do
(Benet 2.5, cited in Kroch & Taylor 1997: 315)
The Loss of V2 in the History of English 31
There are a number of accounts for the loss of the V2 constraint available in the
literature. For example, van Kemenade (1987) explains this loss as a result of a change in
the position of Infl. According to her, in Old and Early Middle English, Infl is outside the
sentential node “S”. As such, in order to be licensed it must be lexicalized either by a
complementizer or by V2 movement. In late Middle English, Infl moves inside the “S”
node where it is after the subject position. Due to this change in the phrase structure, we
no longer observe the V2 pattern.
According to van Kemenade (1997), the asymmetry between pronominal and full
NP subjects was lost so that “fronting of XP to Spec, CP no longer went hand in hand
with V/I to C” (van Kemenade (1997:349). With no positive evidence for the existence of
verb movement to C, children would fail to acquire the V2 constraint thus causing it to
disappear.
The account given in Fischer et al. (2000) equates the loss of V2 with the loss of
verb movement to F, which then eliminates the pronoun/full NP asymmetry since all the
subjects now come before the verb. They suggest that this loss may be linked to loss of
agreement morphology, but they admit that the motivation behind the loss of V to F
movement is still unclear.
Haeberli (2000) attributes the loss of V2 to the loss of empty expletives.
According to him, empty expletives are only licensed if singular present tense forms are
distinguished from the infinitive. The erosion of the infinitive ending in late Middle
English caused the first person singular to be identical to the infinitive. The empty
expletive could no longer appear in SPEC, Agr; therefore, post verbal subjects could no
longer be licensed.
The Loss of V2 in the History of English 32
By far the most elaborated account of the loss of V2 in late Middle English is the
one given by Kroch & Taylor (1997) (K&T) and Kroch, Taylor & Ringe (2000) (KTR).
They develop a complex scenario of language contact and bilingualism based on the
“double base” hypothesis and the notions of “grammar competition” discussed in section
1. According to these authors, the loss of the V2 constraint was a result of language
contact between Northern English speakers and Scandinavian settlers on the one hand,
and Northern and Southern English speakers on the other.
Between the 9th and 11th centuries, the Vikings (speakers of Old Norse) invaded
and settled the North of England. K&T and KTR hypothesize that, during this time
period, there was a great deal of Old English/Old Norse bilingualism, the Norse speakers
having to learn English (as adults) to properly integrate themselves into the community.
Due to imperfect acquisition by the Vikings, some inflectional endings were levelled in
the North of England. Therefore, verb movement to AgrS to attach inflectional endings to
the verb would no longer take place. If the Old English V2 constraint was satisfied by the
verb in AgrS, it could no longer be so when the verb ceased to move there. Since Old
Norse was also a V2 language, the evidence for V2 in a mixed environment was too
robust to allow the constraint to be lost; thus, in order to preserve V2, Northern Early
Middle English gained verb movement to C.
According to K&T and KTR, the loss of V2 was a result of contact between the
CPV2 northern speakers and the IPV2 southern speakers. Since, in the North, both
pronouns and full NPs invert with non-subject topics, when northerners hear southerners
say a sentence with a preverbal pronoun, they conclude that the sentence comes from a
non-V2 grammar. The northerners would then acquire this non-V2 grammar, which
The Loss of V2 in the History of English 33
would begin to compete with their original V2 grammar. The “non-V2” parameter setting
being the typologically less marked, the V2-less northern grammar would have an
advantage in the grammar competition, and would therefore spread at the expense of the
V2 grammars. It is also possible that there was a southern dialect which had
independently lost V2 (see Prasad (2000) for discussion). This could have favoured the
eradication of V2 in England as a whole.
5. Conclusion
In this paper, we have reviewed the literature on the verb second phenomenon in
the history of English. As shown by the variety of analyses presented in the literature, the
loss of V2 in the history of English could have come about, from a syntactic point of
view, in many different ways. It is for this reason that pure technical analyses are
insufficient to choose between the many logical possibilities that syntactic description
affords. The answer must therefore lie in research into language acquisition and
universals, as well as into the sociolinguistic situation of 14th/15th century England. The
research undertaken by Kroch & Taylor (1997) and Kroch, Taylor & Ringe (2000)
(regardless of any particular syntactic analysis presented in their work) is a first step on
what we believe to be the right path.
The Loss of V2 in the History of English 35
Section 2: Article Summaries
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
Prasad,Rashmi(2000)
• dialectsofMiddleEnglish
• historyofEnglish
• V2 was lostfromsouthernMidE in 2differentways:
a) easternmidlands:lost due tocontactwithnortherndialects
b) extremesouth &westmidlands:independent syntactictriggers(subjectmovement,decliticization ofpronouns)
• texts from1350-1500 :regions showdifferentpatterns of useof V2 over time
a) no drop inusage of V2during thisperiod
b) sentenceswithpreposedelement &full NPsubject: dropin V2 around1400sentenceswithpreposedelement &pronounsubject: nochange orslightincrease inV2
• Kroch &Taylor(1997);Kroch,Taylor &Ringe(1997)
• wrt natureof V2constraintinOld/EarlyMiddleEnglish:Vikner(1991),Iatridou &Kroch(1992),Pintzuk(1991),Kroch &Taylor(1997),Kemenade(1987)
• refinementof analysisin Kroch,Taylor &Ringe(1997)
The Loss of V2 in the History of English 36
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
subject: nochange orslightincrease inV2
Stockwell,Robert P.(1984)
• historyofEnglish
• V2 inEnglishbecamelimited tocertaincontexts
a) late OE:optionalnorm
b) ModE:reflexes ofearlier V2in certaincontexts
a) 70-80% ofmain clauseshave V2
b) some frontedconstituentsobligatorilytrigger V2(wh-constituents,negative &affective adv,AdjP, etc),someoptionally(quotations,someadverbials),somedisallow it(NPs, bareinfinitives)
a) Andrew(1934),Bean(1983)
b) Green(1980)
Swieczkowski(1962)
The Loss of V2 in the History of English 37
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
• V2 withfrontedprogressiveor passiveVP, AdjP, orpredicativePP really acase ofequationalinversion
adverbials),somedisallow it(NPs, bareinfinitives)
• only V2 withbe and a fewother verbs
• participlesmay havebeenreanalyzedasadjectives
Swan,Toril(1994)
• Syntaxof OldEnglish
• Old Englishnot a true V2language,unlike OldNorse
• With non-sentenceadverbials,always AVS inON, butalternationbetween AVS& ASV in OE;with sentenceadverbials(speakercomment; not awell-developedclass in ON),usually ASV inOE, alwaysAVS in ON
• OEgrammarincludesseveralconcurrentword-orderoptions
• wrt ON:Faarlund(1990),Christoffersen (1980,1994)
• wrt OEword order:Stockwell(1984),Stockwell& Minkova(1991),Traugott(1972),Vennemann(1974)
• wrt OE asunderlyingly SOV:Lightfoot(1977),Kemenade(1987)
• wrtdistinctionbetweenpronounsand full NPsubjects inOE:Kemenade(1987)
The Loss of V2 in the History of English 38
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
class in ON),usually ASV inOE, alwaysAVS in ON
Traugott(1972),Vennemann(1974)
• wrt natureof sentenceadverbials:Halliday(1970)
OE:Kemenade(1987)
Haeberli,Eric(2000)
• morphology-syntaxinterface
• historyofEnglish
• syntaxof OldEnglish
• loss of V2due to loss ofemptyexpletives
• evidence frommodernGermaniclanguages:adjuncts mayintervenebetweenfronted V &subject iffsubject is a fullDP, only inlanguageswhich licenseexpletive pro
loss of V2coincides withloss ofexpletive pro-drop in English
• explanation forremainingV2 ordersafter emptyexpletiveslost?Remnants,or a V-to-Coption
• wrtobservedwordorders:Kemenade(1987),Pintzuk(1991),Koopman(1998),Kroch &Taylor(1997),Baekken(1998)
• wrt natureof V2 inOE:Cardinaletti& Roberts(1991),Hulk &Kemenade(1997),Kroch &Taylor
• wrtreasons forloss of V2:Fuss(1998),Kemenade(1987),Platzack(1995),Roberts(1993),Kroch etal. (2000),Lightfoot(1997)
The Loss of V2 in the History of English 39
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
a) decliticization ofsubjectpronounsdoesn’texplain it
• emptyexpletiveslost becauseof changes inverbalmorphology
drop in Englisha) pronominal
subjectsbehaving likenon-pronominalsubjectsshould havereinforcedV2; noexplanationfor whypronouns lostclitic status;objectpronouns notdecliticized,contrasts insyntacticbehaviour ofpronominaland non-pronominalsubjectsremain inEarly ModE
• evidence frommodernGermaniclanguages:emptyexpletives onlylicensed inlanguageswhere all sgpres verb forms
• loss of V2is asyntacticeffect of amorphological change
Cardinaletti& Roberts(1991),Hulk &Kemenade(1997),Kroch &Taylor(1997),Pintzuk(1991,1993)
The Loss of V2 in the History of English 40
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
lost becauseof changes inverbalmorphology
• V-frontingtargets Cwhen anoperator isfronted and alower headwhen a non-operator isfronted; 2subjpositions
Germaniclanguages:emptyexpletives onlylicensed inlanguageswhere all sgpres verb formshave clearagreementmorphology
assimilationof 1sg &infinitive formsin Engcoincided withloss of emptyexpletives
exceptionalMidE text withhigh frequencyof V2 also hashigh frequencyof distinctinfinitive form
• distribution ofV2 in OE: withnon-operatorfronted,impossible withpronominalsubjects &optional withnon-pronominalsubjects; with
syntacticeffect of amorphological change
The Loss of V2 in the History of English 41
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
operator isfronted and alower headwhen a non-operator isfronted; 2subjpositionspossible, oneabove & onebelow thelower target;pronounsmust occurabove butfull DPsubjects canstay below ifhigherpositionfilled byemptyexpletive
fronted,impossible withpronominalsubjects &optional withnon-pronominalsubjects; withoperatorfronted, occurswith both typesof subjects
Koopman,Willem F.(1997)
• syntaxof OldEnglish
• V2 effectmuch lessstrong in OEthan in Dutch& German
• anyconstituent canbe topicalized,but differentconstituentscause varyingfrequencies ofinversion
The Loss of V2 in the History of English 42
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
cause varyingfrequencies ofinversion
Fischer,Olga;Kemenade, Ans van;Koopman,Willem;Wurff,Wim vander (2000)ch 4
• syntax(synchronic &diachronic)
• IP consistsof FP (hostspronouns),NegP, TP;Vf lands in Cfor questions,neg-initial,(th)a-initial,but in F(withpronouns inspecF) intopic-initial
• FP hasmorphologyfor subj-Vagreement;V-to-Fmovementlost alongwithagreementinflection
• root questionshave VS order,embeddedquestionsdon’t;embeddedtopicalizationallowed only inquoted rootclauses &unaccusativeconstructions;contrastbetweenpositions ofpronominal andnon-pronominalsubjects (esp.in negative-initialsentences)
• V2 withtopicalizationlost at the sametime as someverb-agreementinflections (sg-pl constrasts)
• Learnersdrawconclusionsabout wordorder basedoninflectionalmorphology ratherthan onobserved
• wrt dialectvariation inMidE:Kroch &Taylor(1997)
Kemenade(1987),Pintzuk(1991)
The Loss of V2 in the History of English 43
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
agreement;V-to-Fmovementlost alongwithagreementinflection
time as someverb-agreementinflections (sg-pl constrasts)
about wordorder basedoninflectionalmorphology ratherthan onobservedword orderpatterns
Cardinaletti, Anna;Roberts,Ian (1991)
• Nominative-Caseassignment
• Syntaxof OldFrench
• Syntaxof OldEnglish
• somelanguageshave twosubjectAgrP’s (onedirectlybelow Comp,one lower);higher oneassignsnominativecase
a) cliticsoccupyAgr1E(head ofhigherAgrP);verbalinflectionin Agr2E
• evidence fromIcelandic (V2in embeddedclauses), OldFrench (V2 inwh-clauses),Yiddish(embeddedtopicalization)
a) evidencefrom German(ordering ofclitics wrteach other &full subject),medievalRomance(Tobler/Mussafia law onposition ofclitics), OldEnglish andOld High
• wrt topotential A-positions:Rizzi(1991)
• wrt Dutchclitics:Zwart(1991)
• wrtRomanceclitics:Mussafia(1983)
• cliticsalways inheadpositions:Kayne(1991)
• wrtanalysis ofTobler-Mussafiaeffects:Alberton(1990),Benincà(1989)
The Loss of V2 in the History of English 44
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
verbalinflectionin Agr2E
b) Agr1E (orAgrE inlanguageswith onlyone)assignsNom underagreementwhere it isa cliticposition, orwhere itcontains aninflected Vwhich iscoindexedwith theNP in thespecifier;possibilityofassigningundergovernmen
medievalRomance(Tobler/Mussafia law onposition ofclitics), OldEnglish andOld HighGerman(interactionof clitics &V2)
b) possiblesubjectpositions inIcelandic,German, OE,OF, Dutch,Italian,English
Kayne(1991)
• wrtNominative-Caseassignment:Koopman& Sportiche(1991)
• wrt nullsubjects:Rizzi(1986)
• wrtconnectionbetween V2& nullsubjects inOF:Vanelli,Renzi &Benincà(1986)
• wrtdistributionofreferentialnullsubjects inOF texts:Hirschbuhler (1990)
The Loss of V2 in the History of English 45
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
with theNP in thespecifier;possibilityofassigningundergovernment is aparametricchoice
c) prolicensed inconfigurations whereNom canbeassigned,except withSpec-headcoindexation
• OF does nothavegeneralizedembeddedtopicalization(does haveStylisticFronting): subjcan be in eitherAgr
c) distributionof pro inIcelandic,German, OF
• embedded V2rare with overtsubjects (exceptin very earlytexts), althoughcommon withnull subjects
subjects inOF texts:Hirschbuhler (1990)
The Loss of V2 in the History of English 46
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
StylisticFronting): subjcan be in eitherAgra) after 13th
c., only oneAgr, whichcannotlicense nullsubj underagreement(inflectionalmorphology toopoor):related toloss ofTobler-Mussafiaeffect &morphological case
• OE cliticsare in Agr1E;V can be inAgr1E or CEin mainclauses, Agr2Ein embeddedclauses
null subjects
a) in embeddedclauses,referentialnull subjectsnot seen inprose after12th c; clitic-first orders inyes/noquestionsappear in 13th
c prose; OFcase systemlost between12th & 14th c;amongGermaniclang,morphological casecorrelateswith doubleAgr
• distribution ofclitic pronounswrt V and subject
The Loss of V2 in the History of English 47
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
are in Agr1E;V can be inAgr1E or CEin mainclauses, Agr2Ein embeddedclauses
clitic pronounswrt V and subject
Kohonen,Viljo(1978)Ch 5
• syntaxof OldEnglish
• historyofEnglish
• relationshipsbetweenwordorderphenomena
• shift fromSOV to SVOpartly due totendency toplace newinformationafter V
• topicalization:
a) triggeredbyconnectionwithprecedingdiscourse,emphasis;affected bydistinctionbetweenobligatory&peripheralelements,weight ofconstituents, stylisticfactors
• quantitativedata (OE): newinformationmore oftenafter V thangiven
•
a) giveninformationmost oftenfronted;fronting oflong or newconstituents,negators,verbs givesemphaticeffect;consecutivesentencesmay havesimilarpattern
• Lightfoot(1977),Johansson(1968),Thompson(1978)
• wrthierarchiesofgivenness:Chafe(1970),Sgall et al.(1973)
• wrttopicalization independentclauses:Hooper &Thompson(1973),Green(1976)
The Loss of V2 in the History of English 48
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
peripheralelements,weight ofconstituents, stylisticfactors
b) inversionused withlong/newsubjects tosolveconflictbetweeninformation andgrammaticalstructures;XSV orderusual withshort,givensubjects
c) when SVXorder fixedfor longersubjects,topicalization blocked
sentencesmay havesimilarpattern
b) VS ordermorecommon withlongersubjects andwith newsubjects
c) proportion oftopicalizationshoulddecreaseas wordorderbecomesfixed
(1973),Green(1976)
• wrt triggersoftopicalization:Johansson(1968),Enkvist(1975,1976)
• wrtdevelopment ofpassives:Lightfoot(1977)
The Loss of V2 in the History of English 49
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
on blocked
• development ofpassivizationconnected tochange fromSOV to SVO
• extraposition of clausalsubjects/objects: insertion ofdummy subjectdevelopedalong withSVX syntax
• Subjectlesspassives declinebetween 10th &13th c
• hit and thaeroptional, butincreases from10th to 13th c
decreaseas wordorderbecomesfixed
• iftopicalizationalso relatedto SVOorder,passivizationshouldincrease astopicalizationdecreases
Hulk,Aafke;Kemenade, Ans van(1995)
• syntax:verbsecond,pro-drop
• historyofEnglish
• languagesmay be C- orI-oriented;dominanthead must belexicalised
• V2 languageshave V-to-Cwhen nocomplementizer in C
• Lightfoot(1979)
• Wrt OF:Adams(1987)
• Wrt OF asIV2:Hirschbuhler &Junker(1988),Lemieux &Dupuis(1995)
The Loss of V2 in the History of English 50
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
ofEnglish
• historyofFrench
lexicaliseda) I licensed
by havingϕ-featuresidentified(syntactically ormorpologically); Vcan assignCase iffgovernedby fullylicensed I
b) pro-drop:pro mustbegovernedbydesignatedCase-assigninghead (C orI); contentmust beidentified
• OF: C-orientedlanguage;referential prolicensed whenV+I moves toC
a) impersonalsentences: Inot fullylicensed, soV cannotassign Case(but P ormorphologycan)
b) cross-linguisticdistributionof full andexpletivepro-drop
• referential pro-drop only in rootclauses
• WrtembeddedV2 in OF:Vance(1988)
(1988),Lemieux &Dupuis(1995)
The Loss of V2 in the History of English 51
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
language;referential prolicensed whenV+I moves toC• MidF:increase of V3structures,embedded V2a) cliticizatio
n of subjectpronouns:XP-pronoun-VfambiguousbetweenCP and IPinterpretation; thisleads toshift fromC-orientationto I-orientation
b) later I losesmorphological ϕ-features;subj NP inSpec,IPbecomesobligatory
clauses
• V2 becomesallowed incomplements ofnon-bridge Va) V1 structures
& embeddedV2(characteristics of IV2lang) becomemorecommon;pro-dropavailable inroot andnonrootcontexts
b) V2 & pro-dropdisappear
The Loss of V2 in the History of English 52
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
cal ϕ-features;subj NP inSpec,IPbecomesobligatory
• OE: C-oriented langa) pronominal
subjectsusuallybefore V;whendiscrepancy betweennominalandpronominalsubjectslost,nominalsubjectsusepronominalpattern inmanydialects
b) as V2declines,Engchangesfrom C-oriented toI-oriented
disappear
• Root/nonrootasymmetrya) data from
Wycliffitesermons(MidE)
b) loss ofexpletive pro-drop
The Loss of V2 in the History of English 53
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
declines,Engchangesfrom C-oriented toI-oriented
• I has ϕ-features in OFbut not in OE
expletive pro-drop
• distribution ofreferential pro-drop, subjects
• Eng:Spec,IPobligatorilysubjectposition;cannot beIV2
Platzack,Christer(1995)
• Syntax:verbsecond,pro-drop
• HistoryofEnglish
• HistoryofFrench
• English &French lostV2 becausethey hadSVO &subjectclitics:unambiguousV2 sentenceswere rare
• NominalAgr(connected toagreementmorphology,V-to-I & pro-drop) lost atthe sametime as V2 inEng, but later
• ME:pronominalsubjects keepsame patternsafter objectclitics lost;inversionoptional in 14th
cMF: subjectpronounsbecome clitics
• ME: subject-Vagreementweakened in14th c; 14th csentences withdo have tenseon both verbs(Infl-lowering);pro-dropdeclines
• wrttheoreticalframework:Holmberg& Platzack(1988,1995(?)),Platzack &Holmberg(1989)
• wrt reasonsfor loss ofV2:Kemenade(1987),Adams(1987)
• wrt loss ofV2: Vance(1989),Weerman(1989)
• theoreticalframework:V2 langhavefiniteness([±F])feature in C,non-V2 langin I
a) [+F] mustgovernNominative;Nominative must begovernedby headhosting[+F] or bychainwhosehead is
The Loss of V2 in the History of English 54
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
(connected toagreementmorphology,V-to-I & pro-drop) lost atthe sametime as V2 inEng, but laterin Fr
weakened in14th c; 14th csentences withdo have tenseon both verbs(Infl-lowering);pro-dropdeclinesMF: pro-dropextended tomore contextsafter loss of V2
(1987),Adams(1987)
• wrt loss ofV-to-I inEng:Pollock(1989)
e must begovernedby headhosting[+F] or bychainwhosehead isgovernedby [+F]
b) Agr(hosted inI) may benominalor not;nominalAgr bearsNominative Case &allowspro-drop
Kemenade, Ans van(1987)
• Syntax:verbsecond
• HistoryofEnglish
• Old EnglishunderlyinglySOV (θ-rolesassignedfrom right toleft)
• Data from OEcompared withDutch:distribution ofV and particles
• Movementrulesneeded:extraposition of objectsto right ofV, V2, V-raising toform verbalclusters
• wrt theoryof languagechange:Lightfoot(1979)
• wrt wordorder inOE:Lightfoot(1976,1977),Canale(1978),Koopman
• wrt wordorder inOE: Malsh(1976),Reddick(1982)
• first twopoints fromch. 2, therest from ch.6
The Loss of V2 in the History of English 55
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
• In V2 lang,INFL is inCOMP &triggersmove V;head of S isV in V2 langand INFL inother lang;topicalizationis optional
• 2 majorchanges inMidE:
a) changefrom SOVto SVO(change indirection ofθ-marking,ca 1200),because ofcommonsurface VOpatterns &loss of casemorphology
• Evidence fromDutch, WestFlemish
• V2 ordersfound afterchange to VOorder complete
a) Position ofparticlesrelative to V;V-raising &extrapositiongive surfaceVO order,loss ofoblique(dat/gen)case & lossof clitics onVP date to ca1200
V, V2, V-raising toform verbalclusters
order inOE:Lightfoot(1976,1977),Canale(1978),Koopman(1984)
• wrtdiagnosticsfor SOVorder:Koster(1975)
• wrtextraposition:Stockwell(1977),Kuno(1974),Pintzuk &Kroch(1985)
• wrtstructure ofV2: Evers(1981,1982),Cremers &Sassen(1983),Platzack(1983),Koopman
• wrtstructure ofV2: Travis(1984),Weerman(1986),Safir(1982)
The Loss of V2 in the History of English 56
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
surface VOpatterns &loss of casemorphology
b) loss of V2(ca 1400),because ofdifficultyininterpreting subject-initialsentencesas V2 &reductionin verbalagreementmorphology
(dat/gen)case & lossof clitics onVP date to ca1200
b) Decrease insubj-Vinversion inthe early 15th
c, changes incomplementizers; subjectpronounscease tobehave asclitics
b) INFLcomes toassign casefrom right toleft, oppositeto V & P atthis stage
(1981,1982),Cremers &Sassen(1983),Platzack(1983),Koopman(1984), deHaan &Weerman(1985),Taraldsen(1983)
• wrt clitics:Borer(1983),Aoun(1985)
• wrt changefrom SOVto SVO:Canale(1978),Hiltunen(1983)
Anderson,Stephen R.(2000)
• optimalitytheory
• placementof specialclitics basedonmorphological processes:clitics areaffixes on thephrase
• evidence fromHittite, Serbo-Croatian,Macedonian,Kwakw’ala,French,Sanskrit
• wrtdistinctionbetweensimple(phonologicallydependent)clitics &special(syntacticall
• wrtMacedonian: Tomi_(1996)
The Loss of V2 in the History of English 57
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
(2000) theory• morpho
syntax:clitics,verbsecond
clitics basedonmorphological processes:clitics areaffixes on thephrase
• clitics areintroducedsimultaneously (ratherthan by a setof orderedrules) &arrangedbased onelement-specificconstraints(somestronger thanothers):edgemost,non-initial,integrity(preventsword orphrase frombeinginterrupted)
Croatian,Macedonian,Kwakw’ala,French,Sanskrit
• evidence fromSerbo-Croatian,Italian,Bulgarian,Warlpiri,EuropeanPortuguese
betweensimple(phonologicallydependent)clitics &special(syntactically unusual)clitics:Zwicky(1977)
• wrt cliticsasmorphology: Anderson(1992,1993, 1995,1996)
• wrt Hittite:Garrett(1990)
• wrt Serbo-Croatian:Browne(1974),Bo_kovi_(1995)
n: Tomi_(1996)
• wrtstylisticfronting:Maling(1980),Jónsson(1991)
The Loss of V2 in the History of English 58
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
beinginterrupted)
• V2 based onthe sameconstraints:non-initial(Vfin,S)&edgemost(Vfi
n,L,S) causesthe verbmovement
• evidence fromIcelandic • V need not
always move:V2 can besatisfied bysubject in itsnormalposition or bystylisticfronting
• wrtWarlpiri:Simpson(1991)
Roberts,Ian (1993)
• syntax:verbsecond,nullsubjects
• historyofFrench
• historyofEnglish
• languageacquisition
• in V2contexts, Cmorphologically selectsAgr toincorporatewith it; headcontainingAgr musthave filledspec
• Fr lostoption toassignNominativeundergovernment
• evidence fromGerman (pro-drop, clitics),West Flemish(complementizer agreement),Dutch (clitics)
• distribution ofsimple &complexinversion
• possibilityof V2 &nullsubjectsdepends onsetting ofNominative-assignmentparameter
• wrtstructure ofI: Pollock(1989),Cardinaletti& Roberts(1991)
• wrt Case-assignment:Sportiche(1988),Koopmann& Sportiche(1990)
• wrtincorporation: Baker(1988),Rizzi &Roberts(1989)
• wrt tostructure ofOF V2:Adams(1988),Dupuis(1989)
• wrt pro-drop inOF:Adams(1987),Dupuis(1988,1989)
• wrt MidFsubjectpronounsas clitics:Adams(1987)
• first pointfrom ch. 1,next 4 fromch. 2, lastpoint from3.4
The Loss of V2 in the History of English 59
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
governmenta) OF null
subjectswerelicensedundergovernment
b) V2 lostbecauseNominative becameonlyassignedbyagreement(subjecthad to bebefore V)
• OForiginallyhad 2 AgrPs,V2movement tothe higherone; Agr1lost in 13th c.,V2 becamemovement toC
a) distributionof nullsubjects
b) distributionof SV andVS orders
• distribution ofV2, freeinversion, nullsubjects
setting ofNominative-assignmentparameter
n: Baker(1988),Rizzi &Roberts(1989)
• wrtinterrogative inversion:Rizzi(1990)
• wrtlicensing ofpro : Rizzi(1986),Jaeggli &Safir (1989)
• wrtstylisticfronting inOF: Dupuis(1989),Lema &Rivero(1990)
• wrt leftdislocationin OF:Foulet(1921),Schulze
subjectpronounsas clitics:Adams(1987)
The Loss of V2 in the History of English 60
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
movement toC
• MidF:a) more
elementsbecameable toappearinitiallywithouttriggeringinversion;CPreanalysedas AgrP
b) Agr able tolicense prounderagreementwhen C notpresent
• Least EffortStrategy inacquisition:sentencesinterpreted tocontainshortestpossiblechains
a) evidencefrom MidF,Swedish,Icelandic: Vmovement,possibility ofV3 orders
b) distributionof pro
(1921),Schulze
• wrt nullsubjects inMidF:Vance(1989),Hirschbuhler (1991)
• SubsetPrinciple inacquisition:Berwick(1985)
• Wrt V2 inOE:Kemenade(1987)
The Loss of V2 in the History of English 61
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
possiblechains
• MidE lostV2 because itacquiredNominative-assignmentunderagreement,after subjectpronounsdecliticized
Lightfoot,David(1999) ch.6
• languageacquisition &change
• historyofEnglish
• childrenscanenvironment(simplesentences)fordesignatedcues on howto setparameters
• V2sentencesbegin witharbitrary XPin Spec,CP;material inSpec,CPmust belicensed bylexical C
• evidence fromCreoles, signlanguage
• model ofacquisition:Dresher &Kaye(1990),Dresher(1998),Fodor(1998)
• wrt dialectsof MidE:Kroch &Taylor(1997)
• wrt signlanguage:Goldin-Meadow &Mylander(1990),Newport(1998),Supalla(1990)
• models ofacquisition: Gibson &Wexler(1994),Clark(1992)
• wrt loss ofV2: Clark& Roberts(1993)
The Loss of V2 in the History of English 62
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
Spec,CPmust belicensed bylexical C
a) cue for V2issentencesbeginningwithnonsubjects; V2acquired ifcue isrobust
b) twodialects ofMidE, onewith V2and onewithout;cue is notrobust forchildrenwho hearboth
b) lowerpercentage ofnonsubject-initialV2 clauses inMidE than inmodern V2 lang,V3 orders withpronouns
Goldin-Meadow &Mylander(1990),Newport(1998),Supalla(1990)
Pintzuk,Susan(1999)
• syntaxof OldEnglish
• Kemenade’s(1987)analysisdoesn’t work
• arguments forverb projectionraising don’twork withdifferent deepstructure;clitics assumedto behavedifferently insubordinate &main clauses;
• wrtproblemswith radicalreanalysis:Allen(1990)
• wrt syntaxof OE:Kemenade(1987)
• first pointfrom ch. 2,next threefrom ch. 3,last pointfrom ch. 6
The Loss of V2 in the History of English 63
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
• language change
doesn’t work
• Clauses varybetweenINFL-medial& INFL-final
work withdifferent deepstructure;clitics assumedto behavedifferently insubordinate &main clauses;not all surfaceverb positionsaccounted for;frequency ofverb-medialorder in main& subordinateclausesanalysed asunrelated,although theyincrease at thesame rate
• distribution ofparticles, objectpronouns,monosyllabicadverbs insubordinateclauses;frequency ofverb raising inINFL-finalclauses
reanalysis:Allen(1990)
• wrt syntaxof OE:Kiparsky(1995)
• wrtnominativevs neutralAGR:Santorini(1992),Platzack &Holmberg(1990)
(1987) from ch. 3,last pointfrom ch. 6
The Loss of V2 in the History of English 64
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
• Double basehypothesis:variation inthegrammars ofindividualspeakers
• V2 in OEusuallymovement toclause-medial INFL
a) V moves toCOMP inquestions,V1 clauses,& clausesbeginningwithcertainadverbs
• V2 lostbecauseAGRchangedfrominherentlynominativeto neutral (socame to needto be c-commanded
frequency ofverb raising inINFL-finalclauses
• variationwithin texts byindividualauthors
• distribution ofclitics
The Loss of V2 in the History of English 65
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
changedfrominherentlynominativeto neutral (socame to needto be c-commandedby a thematicantecedent)
Kroch,Anthony;Taylor,Ann;Ringe,Donald(1997)
• dialectsofMiddleEnglish
• languagecontact
• historyofEnglish
• 2 types ofborrowing:ordinaryborrowing(of contentwords from2nd languageinto 1st) &imposition(borrowingofgrammaticalfeatures from1st languageinto 2nd)
• V2 lostbecause ofdialectcontact:
• evidence fromcode-switchingin nativeYiddishspeakers withfluent English;Norse functionwords inEnglish
• wrt 2 typesof V2languages:Diesing(1990),Pintzuk(1991),Santorini(1992)
• wrtScandinavian influencein England:Geipel(1971),Stenton(1967),Ekwall(1936),Morse-Gagné(1992,1993)
The Loss of V2 in the History of English 66
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
contact:a) southern
dialectspreserveOE IP-V2pattern(movementto Agr)
b) northerndialects areCP-V2
c) northernspeakersinterpretmanysouthernsentencesas non-V2& acquirecompetinggrammar;in south,IP-V2 firstreplacedwith CP-V2 andthen lost
a) distributionof subjectinversion in13th c.Midlandstexts
b) distributionof subjectinversion in anorthern text
c) comparisonof northern &southernmanuscriptsof one text;inversion inChaucer(southern)
(1936),Morse-Gagné(1992,1993)
• wrtimposition:vanCoetsem(1988),Appel &Muysken(1987),Prince &Pintzuk(1984)
• wrt syntaxof OE :Pintzuk(1991),Haeberli &Haegeman(1995),Kroch &Taylor(1997)
• wrtconnectionbetweenmorphologicalagreementand V-to-Imovement :Platzack &Holmberg
The Loss of V2 in the History of English 67
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
• northerndialectbecomes CP-V2 becauseof Norseinfluence
a) NorsesettlerslearningEnglishsimplifyverbalagreementparadigmdue tophonological issues
b) V-to-Agrmovementlost due toreductionofagreement;verbmovementreanalysedas V-to-C
• placement of(originally null)subjectpronouns inNorthumbrianglosses of Latintext
a) evidencefromNorthumbrian texts
b) evidencefromnegation inMidE andModernScandinavian
betweenmorphologicalagreementand V-to-Imovement :Platzack &Holmberg(1989),Roberts(1993),Rohrbacher(1994)
• wrt splitINFL:Pollock(1989)
The Loss of V2 in the History of English 68
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
Yang,Charles D.(2000)
• modeloflanguageacquisition &change
• variationalapproach:languageacquisition isacompetitionprocessamong apopulation ofgrammars
• languagechange ischange in thedistributionof grammars
a) caused bychange inthelinguisticevidence,not byindividualmislearning
b) if onegrammarhas anadvantageoveranother, itwillovertake it
• research inmathematicalpsychology;gradualness ofacquisition
a) competenceof childrenacquiringlanguage;scale oflanguagechange
b) mathematicalevidence
• wrtcompetition-basedlearning:Atkinson,Bower &Crothers(1965),Bush &Mosteller(1951,1958)
• wrt whatdeterminesoutcome ofacquisition:Clark &Roberts(1993)
• wrtvariation ingrammars:Kroch(1989),Kroch &Taylor(1997),Kroch,Taylor, &Ringe(2000),Pintzuk(1991),Santorini(1992)
• wrt theoryoflearning:Gibson &Wexler(1994)
• wrt finalstate ofacquisition: Clark &Roberts(1993)
The Loss of V2 in the History of English 69
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
overanother, itwillovertake it
• advantage ofV2 lost inFrenchbecause ofnull subjectsand increasein V>2patterns
• in MidE:a) south: with
loss ofmorphological case,cliticsreanalyzedas NPs,increasingtheadvantageof SVOover V2
b) north:contactwith southleads tomixture ofgrammars&weakeningof V2
• distribution ofsubjectinversion &pro-drop inMidFr
• distribution ofinversion withnominal &pronominalsubjects
• sameanalysismay extendto otherWesternRomancelanguages
Taylor, &Ringe(2000),Pintzuk(1991),Santorini(1992)
• wrt todialects ofMidE:Kroch &Taylor(1997)
• wrt linkbetweensubjectclitics &V2:Kemenade(1987)
• wrt linkbetweenclitics &morphological case:Kiparsky(1997)
The Loss of V2 in the History of English 70
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
leads tomixture ofgrammars&weakeningof V2
Koopman,Hilda(1997)
• theoreticalsyntax
• pied-pipingis common,even withheavyconstituents
• linear orderof overtterminalelementscorrespondstoasymmetricc-command
a) projectioncan’t haveterminalelements inboth Specand head;only one isovert
• evidence fromBantu lang.,Eng (DP-internal); Fr(IP); Nweh(complement ofNegP)
a) doubly filledComp filter,DoubleVoice filter,evidencefrom Dutch,Eng, Fr wrtapparentviolations;evidencefrom English,Dutch, Irish
a) someprojections, suchas CP,must bemorearticulated thanpreviously thought
• Kayne(1994),Koopman(1994)
• wrt Nweh:Nkemnji(1995)
• wrt DoubleVoice filter:Sportiche(1992)
• wrtdevelopments ofEnglish :Kemenade(1987), W.Koopman(1990),Roberts(1993)
The Loss of V2 in the History of English 71
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
overt
b) if head &Spec bothovert, onemaybecomesilent
• projectioninterpretableiff activatedby lexicalmaterial:lexicalmaterialmust beassociatedwith eitherSpec or headduringderivation
• ModE hasmore Specmarking thanOE
apparentviolations;evidencefrom English,Dutch, Irish
b) evidencefrom Italian,Catalan,Spanish(negativehead); Dutch& German(topic drop),pro-drop
• evidence fromEnglish (CP indeclaratives,wh-movement)
d thanpreviously thought
The Loss of V2 in the History of English 72
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
more Specmarking thanOE
a) OE(apparent)V2involvedtopicmovementto Spec,CPwith Vstayinglower
b) VS orderhas headmovementof V, SVorder haspied-pipingof an XPcontainingthe V;overtbound headforces headmovement,but lack ofit doesn’tprevent it
b) coexistenceof V movement& pied-piping,rather thansuddenreanalysis withloss of verbalmorphology
a) apparentloss ofV2 reallydue tochangesindistribution oflexicalitemslower inthe clause
The Loss of V2 in the History of English 73
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
but lack ofit doesn’tprevent it
c) weakpronounsnot clitics;preverbalobjectpronounshavemovedindependently,postverbalobjectshave heavyXPmovement
c) evidence fromFr, Bantu clitics
c) need atheory forwhatdetermineswhethersmall XP orheavy XPmoves
Lightfoot,David(1995)
• theoryofacquisition
• historyofEnglish
• acquisitionof V2 shouldbe possiblewithoutaccess tonegativedata:childrenlearn that V2sentencesbegin with anarbitraryphrasalcategory;initialcategoryisn’tcomplement
• evidence fromchildrenlearningGerman,Dutch, English& Frenchinverted forms
• wrt natureof V2:Paardekooper (1971),den Besten(1983),Rizzi(1990)
• wrtacquisitionofauxiliariesin English:Newport,Gleitman &Gleitman(1977),Richards
• wrtmotivationfor V2: denBesten(1983), deHaan &Weerman(1986),Haider(1986),Koopman(1984),Platzack(1986),Safir(1981),Evers(1982),
The Loss of V2 in the History of English 74
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
begin with anarbitraryphrasalcategory;initialcategoryisn’tcomplementso must bespecifier;lexicalspecifierrequireslexical head,so V+Imoves there
• loss of V2:gradualincrease insubject-initial V2sentences,followed byreanalysis &sudden lossof V2
• loss of V-to-I:
• sudden declinein V2 in 14th c.English &French;increase insubject-initialclauses in 11th-12th c. E
ofauxiliariesin English:Newport,Gleitman &Gleitman(1977),Richards(1990)
• wrt loss ofV2 in E:Schmidt(1980),Kemenade(1987)
Koopman(1984),Platzack(1986),Safir(1981),Evers(1982),Taraldsen(1986)
The Loss of V2 in the History of English 75
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
a) loss of V-to-C,verbalinflectionare factorsbut not acompleteexplanation
b) alternativesof dosupport &affixhoppingremovedneed for V-to-I
a) be undergoesV-to-I whenrichlyinflected butnot indialectswhere it’suninflected;V-to-Iretained in E& F after lossof V2,Scandinavianlanguageshave V-to-Iwithoutverbalinflection
b) frequency ofoptional dosupport inlate Middle-early ModernEnglish
Bækken,Bjørg(2000)
• historyofEnglish
• V2 patternspersisted intoEarlyModernEnglish:subject-verbinversiongraduallydecreaseswith most
• Evidencefrom 15th-18th
c. prose texts
• Nevalainen(1997)
The Loss of V2 in the History of English 76
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
(2000) English• influen
ce ofpragmaticfactorsonsyntax
EarlyModernEnglish:subject-verbinversiongraduallydecreaseswith mostinitialelements &increaseswith negativeinitialelements
• inversiondue to endfocus/endweight: newinformationtends tooccur at theend of theclause
c. prose texts
• full NPsubjects invertmore often thanpronouns;distribution ofnew subjects
• wrt initialadverbials:Breivik &Swan(1994),Kohonen(1978)
Bean,Marian C.(1983)
• historyofEnglish
• OE notverb-secondor topic-verb;changeddirectly fromSXV toX’SV
• data from theAnglo-SaxonChronicle &other texts
• wrt theory:Firbas(1964), Li&Thompson(1976),Greenberg(1963),Givón(1977)
• wrttheory:Lehmann(1973),Vennemann (1973),Bartsch &Vennemann (1972),Hawkins(1979),Lakoff(1972),
The Loss of V2 in the History of English 77
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
directly fromSXV toX’SV
a) V2 ordersdue tonarrativestyle
b) preverbalpronominalobjectsgiveappearanceof topic-verbordering
(1976),Greenberg(1963),Givón(1977)
• wrt OE:Smith(1893),McKnight(1897),Saitz(1955),Carlton(1959)
n (1973),Bartsch &Vennemann (1972),Hawkins(1979),Lakoff(1972),Koch(1974),Haiman(1974),Lightfoot(1979),McKnight(1897),Hyman(1975), Li&Thompson(1974)
• wrt OE:Brown(1970),Johansson(1968),Kohonen(1976)
The Loss of V2 in the History of English 78
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
Haukenes,Ella(2004)
• historyofEnglish
• syntax-pragmaticsinterface
• OE notconsistentlyV2: TVX
• ModEnegativeinversion,restrictionson type ofsubject notdirectlydescendedfrom OE
• when SVOorder becamegrammaticalised, V2 ordercame toencodepragmaticinformation
• Data from OE,Old Nordic
• Data fromMidE, earlyModE
• Data fromModE,Norwegian
• Vennemann (1974)
• wrtemotiveprinciple:Firbas(1992)
• wrtpragmaticrenewal:Traugott(1989),Brinton &Stein(1993)
•
de Haas,Nynke(2004)
• historyofEnglish
• languagecontact
• pre-contactNorthumbrian: noloanwords,negationexpressedwithindependentni or ne, nogeneralized–s verbending
• evidence from7th-9th c. poems
• wrt origins& dialectsof OE:Nielsen(1998,1986)
• wrt contactwithVikings:Loyn(1994),Baugh &Cable(1993),Nielsen(1998),
• wrt degreeofcomprehensionbetweenON andOE:Townend(2002),Hansen(1984)
The Loss of V2 in the History of English 79
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
ni or ne, nogeneralized–s verbending
• post-contactNorthumbrian showsevidence ofborrowing,but not ofstructuralinterference
• MidE:southerntexts moreconservative& show lessON influence
a) changes innegationnot due tointerference
• generalized–s ending notdue to ONinfluence
• evidence from8th-11th c texts:moreloanwords, butno borrowedpronouns;doublenegation beginsto appear
• evidence fromMidE texts:texts fromDanelaw showmanyloanwords,borrowedpronouns
a) doublenegationincreasesduring periodof contact
• verbalparadigms ofNorthumbrian/Northern MidEvsMercian/EastMidlands MidE
Vikings:Loyn(1994),Baugh &Cable(1993),Nielsen(1998),Burnley(1992),Townend(2002),Pons-Sanz(2001,2004),Barnes(2000),Ekwall(1930),Samuels(1989)
• wrt dialectsof MidE:Baugh &Cable(1993),Trips(2002)
(2002),Hansen(1984)
• wrtinfluenceof ON:Samuels(1989),Burnley(1992)
• wrt V2:Kroch &Taylor(1997),Trips(2002)
• wrt verbalinflection:Kroch &Taylor(1997),Kroch,Taylor &Ringe(2000),Samuels(1989)
The Loss of V2 in the History of English 80
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
due to ONinfluence
• evidencethat V2changed inthe North,but no clearlink to ONinfluence
Northern MidEvsMercian/EastMidlands MidE
• evidence fromNorthern texts
• wrtchanges inOE & ME:Lass(1992),Fischer et al(2000),Baugh &Cable(1993)
• wrtcontact-inducedchange:Thomason(2001),Sebba(1997),Arends,Muysken &Smith(1995)
• wrtborrowing:Burnley(1992),Townend(2002),Baugh &Cable(1993),Serjeantson(1935)
The Loss of V2 in the History of English 81
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
Baugh &Cable(1993),Serjeantson(1935)
• wrtpronouns:Lass (1992)
• wrtnegation:Kemenade(2000)
• wrt V2:Fischer et al(2000),Kemenade(2000)
• wrt OldNorthumbrian texts:Sweet(1975),Campbell(1959)
• wrt post-contact OEtexts: Kuhn(1965),Skeat(1871),Hogg(2004),Pons-Sanz(2004)
The Loss of V2 in the History of English 82
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
Skeat(1871),Hogg(2004),Pons-Sanz(2004)
• wrt MidEtexts:Morris(1874,1878),Iyeiri(2001),Kock(1902),Kroch &Taylor(1997),Clark(1970),Bennet &Smithers(1968),Fischer et al(2000), Holt(1878),Trips(2002), Hall(1920),Holthausen(1921), Jack(1978),Dance(2003),Stadlmann(1921)
The Loss of V2 in the History of English 83
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
(1920),Holthausen(1921), Jack(1978),Dance(2003),Stadlmann(1921)
Henry,Alison(1997)
• syntactic change
• V2 inimperatives:loss goesthrough stagewhere certainverbs appearto permit V2but don’t
• Individualgrammarsarecategorical,but a pool ofexamplesfromspeakerswithdifferentgrammarswould seemto showgradualdecline
• Evidence fromBelfastEnglish: 3dialects, onewith V2, onewith inversionwithunaccusatives(subject insideVP), one withno inversion
• Speakers ofdifferent ageshave differentdialects
• Restrictedchoice ofgrammars:learnerscheckpossibilitiesagainstdata, ratherthanworkingout rules
• wrtembeddedinversion inBE:McCloskey(1992)
• wrt objectshift:Holmberg(1986,1992),Vikner(1990)
• wrt pro-drop:Burzio(1981),Bouchard(1983)
• wrtnegation inimperatives:Beukema &Coopmans(1989),Zhang(1991)
The Loss of V2 in the History of English 84
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
differentgrammarswould seemto showgradualdecline
negation inimperatives:Beukema &Coopmans(1989),Zhang(1991)
Heycock,Caroline;Kroch,Anthony(1993)
• syntax:movement &licensing
• licensingrelationbetween twoelements issatisfied byhighestposition in achain atwhichlicensingconfigurationoccurs;structureprojectedfrom heads isthe minimumwithin whichall licensingrelations canbe satisfied:unnecessarytraces/projections aredeleted
• evidence fromGerman,Dutch, English(coordination);Eng subjectquestions;subject-initialclauses in V2lang;distribution oftopics inYiddish (andOE)
• wrtGerman:Höhle(1983)
• wrt Dutch:Zwart(1991)
• wrt OEdata:Kemenade(1987,1992),Pintzuk(1991)
• Haider(1988)
• wrt SLF-coordination: VanValin(1986)
• wrtanalysis ofV2: Travis(1984)
The Loss of V2 in the History of English 85
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
traces/projections aredeleted
a) V2 alwaysinvolvesmovementto C, but ifIP isdeleted, Ctakes onfunctionsof I
a) evidencefromGerman, OE,Dutch:distributionof unstressedsubjectpronouns,differentbehaviour ofsubject-initial& non-subject-initialmatrixclauses
Kemenade, Ans van(1993)
• historyofEnglish
• syntax:verbsecond,pro-drop
• lateMidE changefrom C-orientation toI-orientationleads toseveralchanges: lossof V2, loss ofimpersonalconstructionswithoutnominative,loss ofexpletivepro-drop
• evidence fromOE & MidEtexts; decreasein doubly filledCOMP
• wrt data:Schmidt(1981), vander Gaaf(1904),Elmer(1981)
• wrtsentencestructure:Chomsky(1986)
• wrt loss ofimpersonals:Lightfoot(1981),Bennis(1986)
The Loss of V2 in the History of English 86
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
constructionswithoutnominative,loss ofexpletivepro-drop
• languagesmay be C-oriented or I-oriented; C-orientedlang. can’thave full pro-drop becauseC can’tidentifyreferentialpro
• evidence fromDutch, Italian
Chomsky(1986)
• wrt pro-drop:Taraldsen(1978),Chomsky(1982),Rizzi(1982),Platzack(1987)
Kiparsky,Paul(1995)
• historyofEnglish
• in English &earlyGermanic,Spec-C is afocusposition(hosts wh-phrases,demonstratives, negation);can bepreceded bya topicpositionadjoined toCP
• Evidence fromModE, OE, OldIcelandic, OldHigh German:interaction oftopicalization& wh-phrases/negation, placement ofclitic pronouns,word order inconjoinedclauses,differencesbetweenrelative clauseswith a barecomplementize
• ModEmoreconservative than otherGermaniclang. indistinguishing topic &wh landingsites
• wrt OEdata: Allen(1977),Mitchell(1985),Kemenade(1987),Pintzuk &Kroch(1989),Pintzuk(1991)
• wrtorigins ofV2 :Wackernagel (1892),Hock(1982),Anderson(1992),Lenerz(1984,1985)
The Loss of V2 in the History of English 87
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
can bepreceded bya topicpositionadjoined toCP
• Indo-Europeanalso had twoleft-peripheraloperatorpositions, butlacked thecategory ofcomplementizer
a) Germanic(& others)gainedcomplementizers
word order inconjoinedclauses,differencesbetweenrelative clauseswith a barecomplementizer & with ademonstrativepronoun pluscomplementizer
• Evidence fromVedic Sanskrit,Hittite, earlyGreek, Latin:topicalizationwith wh-elements,placement ofclitics, adjoinedrather thansubordinatedrelative clauses
a) earlyGermaniclanguageshavecomplementizers (but OEstill allowsadjoinedrelatives)
Pintzuk(1991)
• wrt OldIcelandic:Maling &Zaenen(1981),Maling(1976)
• wrt OldHighGerman:Behaghel(1932)
• wrt theoryofcliticization: Halpern(1992)
• wrt IEphrasestructure:Hale (1987,1989),Hock(1989),Garrett(1992)
(1984,1985)
The Loss of V2 in the History of English 88
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
complementizers
b) V-to-Cmovementfirsttriggeredbyelements inSpec-C
• OE: V-to-Cmovementobligatory,but Coptionalwhen notrequired
havecomplementizers (but OEstill allowsadjoinedrelatives)
• Evidence fromOE: possibleverb positions
Koopman,Willem F.(1996)
• syntaxof OldEnglish
• pronominalsubjects notalways clitics
• topicalization signalsmain clause;pronominalsubjectsinvert todisambiguate
• distribution ofinversion aftervarious initialelements in OE
• rates ofinversion afteradverbs whichcan also besubordinatingconjunctions:high whenalone, lowerwhen anothertopic present
• • Mitchell(1985)
• wrtsubjectpronounsas clitics:Kemenade(1987),Pintzuk(1991)
The Loss of V2 in the History of English 89
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
subjectsinvert todisambiguate
• difficult toexplainvariableinversion ofnominalsubjects:can’t alwaysbe failure ofV2, left-dislocation
subordinatingconjunctions:high whenalone, lowerwhen anothertopic present
• position ofverbs in non-invertedclauses,distribution ofresumptivepronouns
Kroch,Anthony(2005)
• languageacquisition &change
• historyofEnglish
• once aparameter isset, it can’tbe unset; ifit’s setwrong, thecorrectsetting islearned asanotherdialect
• evidence fromchild languageacquisition:errors in casemarking inGerman &Russian, wh-questions inEnglish, V2 inGerman, tough-movement inEng
• wrt settingsofparameters:Clahsen(1991),Penner(1992)
• wrt FLA:Clahsen(1984),Tracy(1986),Babyonyshev et al(1994),Polinsky (toappear), W& T Labov(1976),Müller
• wrt loss ofV-to-I:Lightfoot(1993,1999)
• draft
The Loss of V2 in the History of English 90
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
• in bothacquisitionerrors &languagechange,speakershavecompetingsettings; overtime, onewins out
• acquisitionerrors maycauselanguagechange byskewinglinguisticdataavailable tochildren
• grammarchangeprecedesfrequencychange
• similar timecourse forcorrection ofacquisitionerrors &historicalchange
• evolution ofYiddish fromGerman (SLA),Icelandic“dativesickness”among children
• evidence fromOE (shift fromOV to VO),Yiddish (shiftto INFL-medial), MidE(loss of V-to-I)\
Babyonyshev et al(1994),Polinsky (toappear), W& T Labov(1976),Müller(1996),Anderson(2001),McKee(1997)
• wrt OEchange toVO order:Pintzuk(1997,2002)
• wrtYiddish:Santorini(1993)
• wrtfrequencyofperiphrasticdo: Ellegård(1953)
The Loss of V2 in the History of English 91
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
• V2 lost fromOE becauseof 3-waycompetitionbetweenNortherndialect,Southerndialect &non-V2dialectcreated bycontact
• evidence from13th-17th c.texts: differentsyntax inNorthern &Southern texts;rates ofinversion fornominal &pronominalsubjects
(1953)• wrt V2 in
OE: Kroch& Taylor(1997),Haeberli(2000),Kemenade(1999,1987),Johnson &Whitton(2002)
Lightfoot,David(1997)
• theoryofacquisition
• number ofparametersshould berestricted to30-40
• degree-0learnability:childrenacquiregrammarbased onmain clauses
• languagechange:parametersare resetbecausetriggerexperienceschange
• mathematicalevidence:number ofpossiblegrammars
• evidence fromDutch &German(acquisition ofOV order),Swedish(deletion of ha)
• evidence fromTongan(ergative casesystem),English (loss ofV-to-I, V2)
• historicaldata giveinsight intowhattriggersparametersettings
• wrtdistinctionbetweenroot &subordinateclauses:Emonds(1976),Ross (1973)
• wrtacquisitioninGermanic:Clahsen &Smolka(1986),Wexler(1994)
• wrtSwedish:Andersson& Dahl(1974)
• ‘subordinate clausestrategy’:Roeper &Weissenborn (1990)
• wrttheories ofchange:Andersen(1973),Weerman(1993),Roberts(1993),Keyser &O’Neil(1985)
The Loss of V2 in the History of English 92
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
are resetbecausetriggerexperienceschange
a) trigger forV2 issentencesbeginningwitharbitraryphrase inSpec,CP(non-subject-initial V2);V2 lostwhen notenoughinitial non-subjects
• no theoryneeded toexplainchanges intriggerexperiences
• IP-V2grammar notlearnable
system),English (loss ofV-to-I, V2)
a) evidencefrom Dutch,German,Norwegian,Swedish,MidE:distributionof initialsubjects &non-subjects
• randomvariationalways present
• evidence fromOE: dataleading to IP-V2 analysis notrobust & notdegree-0
whattriggersparametersettings
(1986),Wexler(1994)
• wrtSwedish:Platzack(1986)
• wrtergativecase:Anderson(1988),Chung(1978)
• wrt MidE:Kroch &Taylor(1997)
(1973),Weerman(1993),Roberts(1993),Keyser &O’Neil(1985)
The Loss of V2 in the History of English 93
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
grammar notlearnable
OE: dataleading to IP-V2 analysis notrobust & notdegree-0
Koopman,Willem F(1995)
• syntaxof OldEnglish
• verb-finalmain clausesallowed inOE
a) stylisticallymarked;perhapsimmaturestyle
b) Latininfluenceonly indirecttranslations
• evidence fromtexts: positionof finite verbs
a) distributionof verb-finalclausesacross textsby sameauthor
b) frequency ofverb-final intranslationscompared toother works
• Pintzuk(1991),Kemenade(1987)
Minkova,Donka;Stockwell,Robert(1992)
• Syntaxof OldEnglish
• verse &prose sharethe samesyntax, butverse hasspecialconventions
a) V2 lessstrict in verse
• evidence fromOE verse &prose texts:frequency ofverb-initialclauses,inversion afterambiguousadverb/subordinator, V2 order
• wrt data:Kuhn(1933),Bean(1983),Denison(1986),Mitchell(1964)
• Campbell(1970)
The Loss of V2 in the History of English 94
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
strict in verse adverb/subordinator, V2 order
(1964)
Vennemann, Theo(1974)
• Theoryof wordorderchange
• principle ofnaturalserialization:languagesare eitherhead final orhead initial;inconsistencies are due tochange
• XVlanguages:
a) disadvantage: requiresubject andobject NPsto beclearlydistinguished
b) advantage:sentenceoperatorsdon’tinterruptthe clause
a) unmarkedSOV &object-focused OSVclauses havesame order ofphrasalcategories;case markingin Japanese,Quechua,Hindi,Persian
b) auxiliariesetc. usuallyafter the mainverb in XVand before itin VX lang
• Behaghel(1923,1932),Trubetzkoy(1939),Greenberg(1966),Bartsch(1972)
• wrtnegation inFrench:Regula(1966)
• wrtreasons forlanguagechange:Lehmann(1973)
The Loss of V2 in the History of English 95
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
operatorsdon’tinterruptthe clause
• languageswith S-OmorphologygenerallyXV; becomeVX whenphonologicalchangeweakensmorphology(first topic-initial, thensubject-initial); ifnewmorphologydevelops,change backto XV
etc. usuallyafter the mainverb in XVand before itin VX lang
• history ofEnglish,Romance lang;verb-finalsubordinates inGerman, OE;sentence braceconstruction inGerman, OE,Early Italian;coexistence ofhead-final andhead-initialpatterns inTVXlanguages;development ofnegation in Fr,Eng, German
Clark,Brady Z(2004)
• theoreticalsyntax
• language change
• changealwaysinvolves aperiod ofvariation;languagechange basedon inherentvariationwithingrammars
• examples ofgradual changein English:raising of shorta, shift fromOV to VO, lossof V-to-I
• wrt natureof change:Weinreichet al (1968),Bender(2001)
• wrt natureof change:Lightfoot(1991),Vincent(2000),Fries &Pike(1949),Kroch(1989),Yang
The Loss of V2 in the History of English 96
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
e change variation;languagechange basedon inherentvariationwithingrammars
• theory:a) LFG:
C(onstituent)-structure(trees),F(unctional)-structure(features),&correspondencerelation
b) OT-LFG:potentialstructuresforsemanticinformationconsideredwrtconstraints;optimalcandidateis chosenbased onranking ofconstraints
a, shift fromOV to VO, lossof V-to-I
• degrees ofungrammaticality, openlexicon
b) example: lossof V-to-I inEnglish
Bender(2001)
• wrtstochasticOT:Boersma(1998),Boersma &Hayes(2001),Bresnan etal (2001)
• wrt LFG:Kaplan &Bresnan(1995),Kaplan(1995),Bresnan(2001),Dalrymple(2001),Kuhn(2001,2003)
• wrt model-theoreticsyntax:Blackburnet al (1993),Blackburn& Gardent(1995),Blackburn& Meyer-Viol (1997),
Vincent(2000),Fries &Pike(1949),Kroch(1989),Yang(2000),Tesar &Smolensky(2000)
• wrt OE:Pintzuk(1999),Kroch &Taylor(2001),Haeberli(2002),Hulk &Kemenade(1997),Kemenade(1999,2000),Fischer etal (2000)
The Loss of V2 in the History of English 97
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
wrtconstraints;optimalcandidateis chosenbased onranking ofconstraintsviolated
c) StochasticOT:constraintsranked onacontinuousscale, withthe rankingperturbedby randomfactors;constraintsassociatedwith arange ofvalues(mayoverlapwith otherconstraints)
c) constraintsmay becategorical inone language& onlypreferencesin another;stablevariation ispossible;stylisticvariation
Blackburnet al (1993),Blackburn& Gardent(1995),Blackburn& Meyer-Viol (1997),Potts (2002)
• wrt ordertheory:Davey &Priestley(1990),Samek-Lodovici &Prince(1999)
• wrtmethodology: Mohanan(2003),Boersma(2001,2004),Keller &Asudeh(2002)
• wrtmethodology:Newmeyer(2003)
The Loss of V2 in the History of English 98
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
constraints)
d) GradualLearningAlgorithm:comparesclauseheard towhatgrammarwouldproduce forassumedunderlyingform; ifmismatched,constraintvalues areadjustedslightly
• 3 wordorders in OE& MidEsubordinateclauses(SOViVf,SVfViO,SVfOVi);all-final mostcommon inOE, all-medial mostcommon inMidE
d) gradualnessof languagelearning
• evidence fromOE & MidEtexts
The Loss of V2 in the History of English 99
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
SVfOVi);all-final mostcommon inOE, all-medial mostcommon inMidE
a) twoalignmentconstraints:HEAD-L(eft) &SPINE-R(alignsheads &theirprojectionsto theright); twostructuralmarkednessconstraints,*S and *I(to choosebetweenstructuresheaded byS orheaded byI)
a) generates allattestedorders;ungrammaticality ofSViOVf (noordering ofconstraintswhere it’soptimal)
The Loss of V2 in the History of English 100
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
S orheaded byI)
b) betweenOE &MidE,SPINE-Rbecamelower-ranked &*S becamehigher-ranked
• OEpronouns arenon-projectingand appear inthe I2
domain; fullNP subjectscan appeareither inSpec,V2 orSpec,I2
a) duringMidE,subjectpronounsbecomeprojecting(due to lossof one ofOE’s 2series ofpronouns)
b) application ofGLA to oneOE and oneMidE text
• evidence fromOE & MidE:distribution ofsubjectinversion,position ofsubjects wrt V2
-adjoinednegativeadverbs,distribution ofmodifiedpronouns, nullexpletives
The Loss of V2 in the History of English 101
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
pronounsbecomeprojecting(due to lossof one ofOE’s 2series ofpronouns)& full NPsubjectsappearmore oftenin Spec,I2
b) 2 structuralmarkednessconstraints,CANON-X’ (c-structurecategoriesmust carrya bar level)& *XP(categoryshould notproject to afullphrase); 4hierarchicalconstraints,SUBJECT(requiresovertsubject in
The Loss of V2 in the History of English 102
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
phrase); 4hierarchicalconstraints,SUBJECT(requiresovertsubject inSpec,I2) &dominanceconstraints*Dom(C,Pronon-
projecting),*Dom(I,Pronon-
projecting),*Dom(V,Pronon-
projecting)(determineposition ofnon-projectingpronouns);2alignmentconstraintsFINITE-L,NEG-L(preventelementsprecedingfinite V,negadverb);
The Loss of V2 in the History of English 103
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
FINITE-L,NEG-L(preventelementsprecedingfinite V,negadverb);*EXPLETIVE
c) CANON-X’ &SUBJECTbecomehigher-ranked;*XP,FINITE-L,*EXPLETIVEbecomelower-ranked
Kiparsky,Paul(1996)
• language change
• syntaxof OldEnglish
• OV changedto VO inGermaniclanguages,where verbfrontingmadeevidenceopaque(enablingcause),because ofgeneralizatio
• evidence fromdirectionalityofcomplementation inGermanic,distribution ofrestricted (mainclause) andgeneral verb-fronting,change inEnglish,
• wrt timecourse ofchange:Pintzuk(1991)
• wrtgrammaticalization:Vennemann (1975),Roberts(1994)
The Loss of V2 in the History of English 104
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
frontingmadeevidenceopaque(enablingcause),because ofgeneralization ofrightwardcomplementation(effectivecause)
• OE:a) Wh-
movementandtopicalization involvedifferentverbpositions
b) topicalization not toSpec,IP; nosuch thingas I-finalstructure
Germanic,distribution ofrestricted (mainclause) andgeneral verb-fronting,change inEnglish,Yiddish, Greek
a) position ofclitics,availability inverb-finalmain clauses,verb-frontinginsubordinateclauses
b) notopicalizationinsubordinateclauses;evidencefrom Dutch& other OVlanguages
a) Kemenade’s(1987)analysisnotcorrect
b) Pintzuk’s(1991)analysisnotcorrect
• wrt Greek:Taylor(1994)
Roberts(1994)
• wrtreanalysis:Stockwell(1977),Lightfoot(1979,1991),Kemenade(1987),Stockwell&Minkova(1991)
• wrt V2 inOE:Kemenade(1987),Pintzuk(1991)
The Loss of V2 in the History of English 105
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
as I-finalstructure
c) Spec,CP isa focusposition;where C ispresent,must belexicallyfilled (bycomplementizer orfiniteverb);where Cfilled bycomplementizer ornotpresent, Vmoves to I;Spec,IP issubjectposition,unless noexternalargument;adjunctionpossible toboth CPand IP
clauses;evidencefrom Dutch& other OVlanguages
c) evidencefrom OE:distributionof preverbalconstituents
correct
The Loss of V2 in the History of English 106
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
externalargument;adjunctionpossible toboth CPand IP
Kemenade, Ans van(1997)
• historyofEnglish
• OE & MidEare CV2
a) pronounsare clitics;cliticization blockedby anoperator infirstposition
b) V2 lost inlate MidEdue to lossofdiscrepancy betweenpronominal& nominalsubjects
• Evidence fromOE, MidE:root/non-rootasymmetry;embeddedtopicalizationonly insubjectlesscontexts, MidEtopics precedenegated Vf
a) position ofpronouns
b) evidencefrom MidE:inversion inChaucer &Wycliffe
• wrt natureof V2: Hulk&Kemenade(1993)
• Pintzuk(1991,1993),Kroch &Taylor(1997)
The Loss of V2 in the History of English 107
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
ofdiscrepancy betweenpronominal& nominalsubjects
Chaucer &Wycliffe
Warner,Anthony(1997)
• language change
• historyofEnglish
• changemotivated byproperties ofUG
• V2 began todecline inearly MidE
• varyingrates ofinversion inMidE may bedue tostylisticfactors ratherthan dialectvariation
• evidence fromloss of V-to-Iin Eng
• evidence fromearly MidE:rates ofinversion,cliticization ofsubjectpronouns
• northern Ruleof St. Benetmay bestylisticallymarked
• loss of V2may be dueto loss of I-finalstructure
• wrt loss ofV-to-I:Lightfoot(1997),Roberts(1993),Kroch(1989)
• wrtimperativeinversion:Henry(1995,1997)
• wrt V2:Kemenade(1997)
• wrt natureofparameters: Zaring &Hirschbühler (1997)
• wrt V2:Kroch &Taylor(1997)
Fontana,Josep M(1997)
• syntax :clitics
• modernRomanceclitics havedifferentpropertiesfrom OldRomance &Germanicclitics; 2Pclitics are
• OR &Germanicclitics mayattachenclitically toalmost any typeof constituent;ModernSpanish cliticsmust attach to
• wrt clitics :Taylor(1990),Halpern(1992)
• wrt OldSpanish:Rivero(1997)
The Loss of V2 in the History of English 108
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
(1997) clitics havedifferentpropertiesfrom OldRomance &Germanicclitics; 2Pclitics arephrasalcategories, inSpec,IP oradjoined toIP or to aprojection inSpec,IP
• Cardinaletti& Roberts(1991) :
a) accountsfor varyingpositionsof clitics
clitics mayattachenclitically toalmost any typeof constituent;ModernSpanish cliticsmust attach toverb
a) evidencefrom German& Dutch(position ofsubject wrtclitics), OF(Tobler-Mussafiaeffects), OE(position ofclitics wrtverb)
(1990),Halpern(1992)
• wrt V2:Diesing(1990),Santorini(1989),Rögnvaldsson &Thráinsson(1990), denBesten(1978)
• doublebasehypothesis:Santorini(1989),Kroch(1994)
• wrt OF asIV2:Dupuis(1989)
Rivero(1997)
The Loss of V2 in the History of English 109
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
b) obscuressimilaritiesbetweenGermanic& ORlanguages
(position ofclitics wrtverb)
b) differinganalyses ofsimilar factsin OE & OF
Kroch, A.(1989)
LanguageChange
• Constantratehypothesis:when onegrammaticaloptionreplacesanother withwhich it isincompetitionacross a setof linguisticcontexts, therate ofreplacement,properlymeasured, isthe same inall of them.
• Loss of V2 inFrench(Fontaine,1985): the rateof loss ofinversion andnull subjects isroughly thesame as theloss of V2order.
• The rate of therise of reprise(leftdislocation) isthe same asthe decline ofV2.
• Grammatical analysisthatdefines thecontexts ofchange isquiteabstract, assynchronicsyntacticstudiessuggest.
• Quantitative studiescanprovidedirectevidenceas to thecausationof change.
Weinreich,Labov &Herzog (1968).
Adams (1987)
Bailey (1973)
Stein (1986)
“The changefrom onegrammar toanother isnecessarilyinstantaneousand its causesare necessarilyexternal” (p.2)“the process oflanguagechange is not afact ofgrammar but afact oflanguage useand so must bestudied withtoolsappropriate tothat domain.”(p.3)
The Loss of V2 in the History of English 110
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
• The set ofcontexts ofchange isnot definedby a sharedsurfaceproperty(like aparticularword ormorpheme),but bysyntacticstructure.
• Linguisticchangesfollow and‘S-shaped’curve.
• The loss ofV2 inFrench wasdue to achange inphrasalaccentwhich forcesproposedconstituentsto movefrom thetopicalization position tothe positionof left-dislocation.
of change. that domain.”(p.3)“Variationoften reflectschoices that arenotcategoricallydetermined bylinguisticprinciples atany level butinstead areonlyprobalisticallyinfluenced byfeatures ofcontaxt andsituation” (p.3)
The Loss of V2 in the History of English 111
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
constituentsto movefrom thetopicalization position tothe positionof left-dislocation.
• ThedislocatedNPs don’tcount for theV2constraint,and assentences ofthis typeincrease infrequency,positiveevidence forV2decreases.
Pintzuk(1993)
V2 inEnglish
• Verbseconding inmain clausesinvolvesverbmovementto Infl ratherthan toComp
• Presence ofV2 insubordinateclauses
• EPP canbe satisfiedthroughverbalmorphology
Iatridou &Kroch (1992)
Santorini(1992)
Verb movesto comp infinite mainclauses:VanKemenade(1987)
The Loss of V2 in the History of English 112
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
movementto Infl ratherthan toComp
• Verbseconding toInfl appliestosubordinateclauses (notonly mainclauses)
• OEexhibitedboth Infl-final andInfl-medialphrasestructure
y (double basehypothesis)
Kemenade(1987)
Kiparsky(1990)
Verb movesto clause-medialfunctionalhead insubordinateclauses:Cardinaletti &Roberts(1991)Haeberli(1991, 1992),Haeberli andHaegeman(1992),Higgins(1991),Koopman(1985),Lightfoot(1991)Tomaselli(1990)
The Loss of V2 in the History of English 113
Article Context Major claims Major Arguments Implications Authorsfollowing
Authorsagainst
Comments
(1990)
The Loss of V2 in the History of English 114
Section 3: Bibliography of V2 in English
Allen, Cynthia. Obsolescence and sudden death in syntax: The decline of verb-final order in early MiddleEnglish. Generative Theory and Corpus Studies: A Dialogue from 10 ICEHL , 3-25. 2000.mouton de gruyter.
Anderson, John. On Variability in Old English Syntax, and Some Consequences Thereof. Transactions ofthe Philological Society 95[1], 9-40. 1997. Malden, MA, Blackwell Publishers.
Anderson, Stephen R. Towards an Optimal Account of Second-Position Phenomena. Optimality Theory:Phonology, Syntax and Acquisition , 302-333. 2000. Oxford U Press.
Baekken, Bjorg. Inversion in Early Modern English. English Studies 5, 393-421. 2000.
Bean, Marian C. The Development of Word Order Patterns in Old English. 1983. Totowa, NJ, Barnes &Noble.
Bobaljik, J. Why Phi? paper presented at the Phi Workshop. 2005.
Burnett, H. and Ferch, E. V2 in the History of French: A review of the Literature. in preparation.
Cardinaletti, Anna and Roberts, Ian. Clause Structure and X-Second. 11/1991. Universita deVenezia//University of Wales.
N. Chomsky. 1995. The Minimalist Program. Cambridge, Mass., MIT Press.
Clark, Brady Z. A Stochastic Optimality Theory Approach to Syntactic Change. 07/2004. StanfordUniversity.
de Haas, Nynke. Dialect Variation and Language Contact: The Influence of Old Norse on MedievalEnglish. 24/12/2004. University of Groningen.
Fischer, Olga, van Kemenade, Ans, Koopman, Willem, and van der Wurff, Wim. The Verb-Secondconstraint and its loss. The Syntax of Early English , 104-137. 2000. New York, CambridgeUniversity Press.
Fontana, Josep M. On the integration of second position phenomena. Parameters of MorphosyntacticChange , 207-249. 1997. New York, Cambridge University Press.
Haeberli, Eric. Inflectional Morphology and the Loss of V2 in English. University of Reading. 09/2000.
Harley, H. Abstracting away from abstract case. Proceedings of NELS 25. 1995. University ofMassachusetts, Amherst.
Haukenes, Ella. `Pragmatic aspects of word order developments in English and Scandinavian languages.NAES 2004 . 2004. Aarhus.
Henry, Alison. Viewing change in progress: the loss of V2 in Hiberno-English imperatives. Parameters ofMorphosyntactic Change . 1997. New York, Cambridge University Press.
Heycock, Caroline and Kroch, Anthony. Verb movement and coordination in a dynamic theory oflicensing. The Linguistic Review 11, 257-283. 1994.
The Loss of V2 in the History of English 115
Heycock, Caroline and Kroch, Anthony. Verb Movement and the Status of Subjects: Implications for theTheory of Licensing. Groninger Arbeiten zur germanistischen Linguistik 36, 75-102. 1993.
Hulk, Aafke. Residual verb-second and the licensing of functional features. Probus 5, 127-154. 1993.
Hulk, Aafke and van Kemenade, Ans. Verb Second, Pro-drop, Functional Projections and LanguageChange. Clause Structure and Language Change , 227-256. 1995. New York, Oxford UniversityPress.
Iatridou, Sabine and Kroch, Anthony. The Licensing of CP-recursion and its Relevance to the GermanicVerb-Second Phenomenon. Working papers in Scandinavian syntax 50. 10/1992 . Trondheim,Linguistics Dept., University of Trondheim.
Kiparsky, Paul. Indo-European Origins of Germanic Syntax. Clause Structure and Language Change , 140-169. 1995. New York, Oxford University Press.
Kiparsky, Paul. The Shift to Head-Initial VP in Germanic. Comparative Germanic Syntax . 1996. Kluwer.
Kohonen, Viljo. Thematic Structure and Transformations. On the Development of English Word Order inReligious Prose around 1000 and 1200 A.D. 138-196. 1978.
Koopman, Hilda. The Doubly filled C filter, the Principle of Projection Activation and Historical Change.1997. UCLA.
Koopman, Willem F. Inversion after single and multiple topics in Old English. Advances in EnglishHistorical Linguistics , 135-150. 1996. Mouton de Gruyter.
Koopman, Willem F. Old English Clitic Pronouns: Some Remarks. Evidence for Old English , 44-87. 1992.Edinburgh, John Donald Publishers.
Koopman, Willem F. Topicalization in Old English and its effects. Some remarks. Language History andLinguistic Modelling: A Festschrift for Jacek Fisiak on his 60th Birthday , 307-321. 1997.Berlin, Mouton de Gruyter. Hickey, Raymond and Puppel, Stanislaw.
Koopman, Willem F. Verb-Final Main Clauses in Old English Prose. Studia Neophilologica 67, 129-144.1995.
Kroch, A. Syntactic Change.Notes: Course notes for LSA139. 2005 LSA Institute. Harvard/MIT.
Kroch, Anthony. If at first you don't succeed: the time course of language acquisition and its implicationsfor language change. presentation at LSA . 2003. University of Pennsylvania.Notes: draft
Kroch, Anthony. Morphosyntactic Variation. Papers from the 30th Regional Meeting of the ChicagoLinguistics Society: Parasession on Variation and Linguistic Theory . 1994 .
Kroch, Anthony. Reflexes of Grammar in Patterns of Language Change. Language Variation and Change 1,199-244. 1989.
Kroch, Anthony and Taylor, Ann. Remarks on the XV/VX Alternation in Early Middle English. 03/1994.University of Pennsylvania.
Kroch, Anthony and Taylor, Ann. Verb movement in Old and Middle English. in Parameters ofMorphosyntactic Change. van Kemenade & Vincent (eds.). Cambridge University Press. 1997
The Loss of V2 in the History of English 116
Kroch, Anthony and Taylor, Ann. Verb-Object Order in Early Middle English. 2000.
Kroch, Anthony, Taylor, Ann, and Ringe, Donald. The Middle English verb-second constraint: a case studyin language contact and language change. 1997. University of Pennsylvania.
Lightfoot, David. Cue-Based Acquisition and Change in Grammars. The Development of Language , 144-177. 1999. Malden, MA, Blackwell Publishing.
Lightfoot, David. How to Set Parameters: Arguments from Language Change. 1991. Cambridge, MA, TheMIT Press.Notes: ch 1, 3
Lightfoot, David. Shifting triggers and diachronic reanalyses. Parameters of Morphosyntactic Change ,253-272. 1997. New York, Cambridge University Press.
Lightfoot, David. Why UG Needs a Learning Theory: Triggering Verb Movement. Clause Structure andLanguage Change , 31-51. 1995. New York, Oxford University Press.
Minkova, Donka and Stockwell, Robert. Poetic Influence on Prose Word Order in Old English. Evidencefor Old English , 142-154. 1992. Edinburgh, John Donald Publishers.
Pintzuk, Susan. Phrase Structures in Competition: Variation and Change in Old English Word Order. 1999.New York, Garland Publishing.Notes: ch. 1, 2, 3.1, 6
Pintzuk, S. Verb seconding in Old English: verb movement to Infl. The Linguistic Review 10 (1993), 5-35
Pintzuk, Susan and Taylor, Ann. Objects in Old English: Why and How Early English Is Not Icelandic.York Papers in Linguistics 2[1], 137-150. 2004. Heslington, York.
Platzack, Christer. The Loss of Verb Second in English and French. Clause Structure and Language Change, 200-226. 1995. New York, Oxford University Press.
Prasad, Rashmi. A Quantitative Analysis of the Loss of V2 in the History of English. WCCFL 19Proceedings , 371-384. 2000. Somerville, MA, Cascadilla Press.
Roberts, Ian. Object Movement and Verb Movement in Early Modern English. Studies in ComparativeGermanic Syntax , 269-284. 1995. Kluwer Academic Publishers.
Roberts, Ian. Verbs and diachronic syntax: a comparative history of English and French . 1993. Dordrecht,Kluwer Academic Publishers.
Roberts, Ian and Roussou, Anna. The History of the Future.
Stockwell, Robert P. On the history of the verb-second rule in English. Historical Syntax , 575-592. 1984.The Hague, Mouton. Fisiak, Jacek.
Swan, Toril. A note on Old English and Old Norse initial adverbials and word-order with special referenceto sentence adverbials. Language change and language structure , 233-270. 1994. Mouton deGruyter. Morck, E., Swan, T., and Westivik, O. J.
van der Wurff, Wim. Deriving object-verb order in late Middle English. Journal of Linguistics 33, 485-509.1997. UK, Cambridge University Press.
van Kemenade, Ans. Syntactic case and morphological case in the history of English. 1987. Dordrecht,
The Loss of V2 in the History of English 117
Foris Publications.
van Kemenade, Ans. Syntactic Changes in Late Middle English. Historical Linguistics 1989 , 235-248.1993.
van Kemenade, Ans. V2 and embedded topicalization in Old and Middle English. Parameters ofMorphosyntactic Change , 326-352. 1997. New York, Cambridge University Press.
Vennemann, Theo. Topics, Subjects, and Word Order: from SXV to SVX via TVX. Historical linguistics I:Proceedings of the First International conference on Historical Linguistics , 339-376. 1974.
Warner, Anthony. The structure of parametric change, and V-movement in the history of English.Parameters of Morphosyntactic Change , 380-393. 1997. New York, Cambridge UniversityPress.
Weerman, Fred. The V2 Conspiracy: A synchronic and a diachronic analysis of verbal positions inGermanic languages. 1989. Providence RI, Foris Publications. Publications in LanguageSciences.
Yang, Charles D. Internal and external forces in language change. Language Variation and Change 12, 231-250. 2000. U.S.A., Cambridge University Press. 2001.