Professional Documents
Culture Documents
July 2015
COURSE NOTES
Note:
Bibliographical references throughout are to the section on ‘References and Bibliography’ at
the end of these notes.
0
© John Charles Smith 2015 Language Change and Historical Linguistics
0. Introduction and suggestions for general reading
The general issues involved in language change are discussed from a rather basic viewpoint in
Aitchison (2000), and, more thoroughly, in a dated but still worthwhile book, by Bynon (1977).
Useful notions can also be gleaned from Hock (1991), Lehmann (1992), and Jeffers & Lehiste
(1979). Three good fairly recent introductions to historical linguistics are McMahon (1994), Trask
(2007), and Campbell (2013). Also interesting is Crowley & Bowern (2010) — written by an
Austronesianist and an Australianist, it eschews the Eurocentric approach of most introductory
works. A more advanced theoretical introduction is Ringe & Eska (2013). Lucid discussion of
sources and methods and the problems associated with them can be found in the following more
advanced works: Lass (1997), especially chapter 2: ‘Written records: evidence and argument’, pp.
44-103; Labov (1994), especially chapter 1: 'The use of the present to explain the past’, pp. 9-27;
and Fleischman (2000). Good specialized dictionaries of historical linguistics are Trask (2000) and
Campbell & Mixco (2007), and relevant terminology can also be found in the standard general
dictionaries of linguistics, such as Matthews (2014) and Crystal (2008), and encyclopædias of
linguistics, such as Frawley (2003), Crystal (2010) and the monumental 14-volume compilation of
Brown (2005).
1. Theoretical Preliminaries
1.1. Two ‘axes’
synchrony (the study of) a phenomenon at one point in time
‘vs.’?
diachrony (the study of) a phenomenon through time
A topical problem... In recent years, diachronic linguistics has returned to prominence, after
almost a century of relative neglect (largely the result of Saussure’s understandable but over-
zealous distinction between synchrony and diachrony in the Cours de linguistique générale,
published posthumously in 1916). This trend has appeared in various guises. Work on
grammaticalization1 and subjectification2 has led to the suggestion that grammars should be viewed
principally as dynamic systems; sociolinguistics has stressed the link between variation and change
and has canvassed ‘the use of the present to explain the past’;3 attention has been drawn to possible
analogies between language change and biological evolution;4 claims about affiliation have led to
major re-evaluations of the aims and techniques of linguistic reconstruction;5 and the application of
contemporary formalisms to language change has not only elucidated the diachronic processes
involved, but has also enlightened us about the formalisms themselves.6 In this, as so often,
linguistics is simply in step with the age; many disciplines, from physics to economics, have also
come to treat phenomena previously considered static as dynamic processes. It’s all very well to
say, as Saussure did, that we can’t study language change properly until we’ve first defined
1
The literature on grammaticalization is vast. Convenient introductions are: Hopper & Traugott (2003) and Lehmann
(2002), available on line at: http://www.christianlehmann.eu/publ/ASSidUE09.pdf . We return to this issue below.
2
See, for instance, Traugott (1982; 1989; 2003); Sweetser (1990). Compare, as an example of subjectification, the
semantic evolution of the English word still, from a spatial item (‘stable in space’: still water, a still day), through
metaphorical (‘space for time’) interpretation as a temporal item (‘stable in time’: He’s still there), to a modal
‘concessive cancellative’ (‘stable in the attitude of the speaker’: Even if it rains, we can still go for a walk). Compare
also the development of English whilst (temporal to adversative), since (temporal to causal), etc.
3
William Labov has used this phrase in many of his publications.
4
See especially Croft (2000); McMahon & McMahon (2012).
5
See, for instance, Ringe, Warnow & Taylor (2002).
6
Compare the following claim by Kiparsky (2008): ‘If language change is constrained by grammatical structure, then
synchronic assumptions have diachronic consequences. Theories of grammar can then in principle contribute to
explaining properties of change, or conversely be falsified by historical evidence.’ (‘Universals constrain change;
change results in typological generalizations.’
1
© John Charles Smith 2015 Language Change and Historical Linguistics
language; but what if change were an essential part of the definition of language...? We can now
appreciate anew the perceptive ‘pre-Saussurian’ observation that Hugo Schuchardt (1842-1927)
made 130 years ago: ‘Jedes Stadium der Sprache ist ein Uebergangsstadium’ [‘Every stage of
language is a transitional stage’] (see Schuchardt 1885).
How languages don’t change (despite the impression given by many books on the history of French
and other languages!) — the ‘metachronic fallacy’:
A → B ✘
In other words, people don't suddenly wake up one morning to find that a language has changed
overnight! Malcolm Bradbury’s fictional Stalinist state of Slaka, in which this sort of thing
happens, is, of course, a parody...
‘But on the third day, when he rises, and goes down to the lobby, to collect, on his way to breakfast, the
red-masted newspaper, he senses that something has changed. Then, over the slow breakfast, he sees
what the change is, a perfectly small one: P’rtyuu Populatuuu has become P’rtyii Populatiii again.’
Malcolm Bradbury, Rates of Exchange (London: Secker & Warburg, 1983)
7
In fact, Chambers & Trudgill (1998), looking at dialect differences, distinguish between pronunciation (neither
systematic nor systemic), phonetics (systematic, but not systemic), and phonology (both systematic and systemic).
Compare: pronunciation, English schedule [ʃedjuəәl] vs. [skedjuəl], but shed [ʃed] vs. *[sked], etc. (just affects one
word, so unsystematic); phonetics Spanish la semana [la semana] vs. [la hemana], English better [betə] vs. [beʔəә]
(systematic, but does not impinge on lexical distinctions, therefore not systemic); phonology Spanish arrollo ([aroʎo]
or [arojo]) vs. arroyo (only ever [arojo]), or cierra ([θjera] or [sjera]) vs. sierra (only ever [sjera]), English putt [pʌt] or
(Northern) [pʊt] vs. put (only ever [pʊt] (systematic and does impinge on lexical distinctions, so also systemic).
2
© John Charles Smith 2015 Language Change and Historical Linguistics
1.5. Two stages of change: actuation and propagation
A change emerges, giving rise to variation (‘actuation’). Subsequently, this change may spread
through the language (‘propagation’).8
An analogy from evolutionary biology: Mutations arise, selection favours certain mutations, which
thrive at the expense of others.
The analogy is not exact, as linguistic ‘mutations’ are in general less random than biological ones
(they can be motivated — see below) and linguistic ‘selection’ generally has a greater element of
randomness (although it may to some extent be socially determined — again, see below).
However, it is thought-provoking.
FURTHER READING: For the relationship between (random) mutation and natural selection in
evolutionary biology, see almost any book by Richard Dawkins — for instance, Climbing Mount
Improbable (London: Viking, 1996; paperback edition, Harmondsworth: Penguin, 1997).
A high-level discussion of the analogies between linguistic change and biological change can be
found in Croft (2000). Croft argues, inter alia, that plant evolution, which allows cross-
fertilization and hybridization (see §1.7.2 below), may be a more relevant analogy for language
change than animal evolution.
diachronic variation variation through time (what historical linguistics is all about)
diatopic variation variation according to place; geographical variation
diastratic variation variation according to social class
diagenic variation variation according to sex/gender
diaphasic variation variation according to register or style
diamesic variation variation according to medium (written vs. spoken, etc.)9
8
On this crucial point, see Weinreich, Herzog & Labov (1968).
9
In practice, the distinction between written and spoken language is often more diaphasic than diamesic.
10
Classic studies of the use of linguistic variables to express or create social identity include Labov (1963), Milroy
(1987), and Eckert (1989; 2000). On overt and covert prestige, see Trudgill (1983a), chapter 4
3
© John Charles Smith 2015 Language Change and Historical Linguistics
‘The U[niformitarian] P[rinciple] holds that we can constrain our hypotheses about the structure and
history of languages of the past only by reference to what we know of contemporary language structures,
linguistic behaviour and changes in progress, since the recoverable information about any language or
speech community of the past is always far more limited than what we can know about languages whose
native speakers we can still observe; and, further, that we can extrapolate into prehistory (and across gaps
in the historical record) only on the basis of what we know from the study of contemporary languages and
the actually documented past. Positing for any time in the past any structure or development inconsistent
with what is known from modern work on living languages is unacceptable, and positing for prehistory
any type of long-term development that we do not observe in documented history is likewise
unacceptable, unless it can be demonstrated that there has been some relevant change in the conditions of
language acquisition or use between the past time in question and later periods which can be observed or
have been documented.’
FURTHER READING: The idea that speakers have a role in language change is not hard to grasp.
On the role of the hearer in language change, see Ohala (1989; 1992; 1993).
The dominant view of sound change in the nineteenth century (echoed in the quotation from
Sturtevant) was that it was exceptionless — that is, a given change took place whenever its
structural description was met, regardless of the particular word involved. This position was
associated with the Neogrammarians (Junggrammatiker), a group of German scholars, of whom the
most famous are probably Karl Brugmann (1849-1919), Berthold Delbrück (1842-1922), Hermann
Osthoff (1847-1909), and Hermann Paul (1846-1921). However, Hugo Schuchardt (1842-1927)
disagreed, and, following pioneering work on linguistic geography (especially the work on the Atlas
linguistique de la France by Gilliéron & Edmont) at the turn of the twentieth century, a rival
hypothesis grew up, according to which sound change took place according to the model of lexical
diffusion — in other words, a sound change took place in some words before others. The ‘slogan’
of the lexical diffusion camp was: ‘Chaque mot a son histoire’ [‘Every word has its [own] history’].
A problem for historical linguists with access to only scant data is that the end result of a sound
change which applies ‘across the board’ at the same time and the end result of a sound change
which is lexically diffuse but which does eventually propagate right through the lexicon are
identical, and the data may not permit the two processes to be distinguished.
4
© John Charles Smith 2015 Language Change and Historical Linguistics
Throughout much of the twentieth century, the lexical diffusion hypothesis was rather airily
dismissed. In fact, there is some evidence that at least some sound changes are lexically diffuse.
Examples of such changes are not hard to find. Extremely frequent items (such as Spanish Vuestra
Merced > Usted, French Mon Sieur > Monsieur) may undergo irregular sound change ([bwestra
meɾθeð] > [usteð], [mɔ̃sjør] > [məsjø]); although ‘one off’ changes are not generally regarded as
examples of lexical diffusion, they do indicate that sound change may affect individual words (this
type of development may be especially frequent when the item in question becomes
grammaticalized: compare Usted, Monsieur, and English do not > don’t, is > ’s, has > ’s, isn’t it? >
innit?, etc.).
The best recent discussion of the Neogrammarian hypothesis vs. lexical diffusion can be found in
William Labov Principles of Linguistic Change: volume 1, internal factors (Oxford: Blackwell,
1994), Part D ‘The regularity controversy’, chapters 15-18, pp. 419-543.
Consider now the following example overleaf of regular changes operating on regular forms to
produce irregularity. Regular and transparent masculine/feminine adjective pairs in Latin (here
given in the accusative case) undergo regular sound change to yield opaque outcomes in French.
As an example of irregular change leading to regularity, we might take the past tense of certain
verbs in English. Strove has become strived for many speakers; throve has become thrived for most
speakers. This change makes the past tense of these verbs more regular — strive, strived and
thrive, thrived, just like arrive, arrived. But it’s not a regular, ‘across the board’ change — there’s
no sign of drove becoming *drived, for instance (possibly because of its high frequency; see
below).
However, the waters appear to be muddied by the fact that, in many varieties of American English,
dived, the regular past tense of the verb dive, has been replaced by the irregular from dove. How
can we account for a variety in which strove becomes strived, yet dived becomes dove? Is this
simply random or chaotic?
Probably not. The arguments here lead us to distinguish between global and local environments.
Globally, most English verbs are regular, forming their past tense in -ed. The shift from strove to
strive is therefore regularization in a global sense. But, if we take English verbs indicating manner
of motion with an /aɪ/ in their stem, we find that, fairly systematically, they form their past tense by
Ablaut or apophony — i.e., by changing their vowel, in this case to /əʊ/: compare drive, drove; ride,
rode; write, wrote; rise, rose, etc. Seen within this local environment, dive becoming dove is a
regularization.
5
© John Charles Smith 2015 Language Change and Historical Linguistics
It is important when discussing frequency to distinguish between types and tokens. Types are
individual items, tokens are occurrences of those items. You might like to think about how many
words there are in the sentence The cat sat on the mat (there are two correct answers!). We
sometimes get different answers according to which class of items we are dealing with. For
instance, the statement that ‘most English verbs are regular’ is almost a truism (that’s what ‘regular’
means) — but only as a statement about types (i.e., most of the verbs listed in the Oxford English
Dictionary, etc. form their past tense and past participle in -ed). If we look at tokens, the situation
changes, and the statement is probably false, because, in an extended sample (say, a million words)
of English text, most of the verb forms are going to be of common verbs like be, have, go, make,
do, etc., all of which are irregular.
In this case, subsequent evidence modifies the hypothesis, and the child comes to learn the form
fought. But in certain circumstances (see the section on analogy above), it may be that the child’s
6
© John Charles Smith 2015 Language Change and Historical Linguistics
hypothesis is not overridden, resulting in an analogical extension of the regular form and a change
in the language — hence, for instance, throve > thrived.
There is a third logical possibility, abduction, which, unlike deduction and induction is logical only
in the sense that it exists — it does not result in a conclusion that can be sustained in logic.
Consider the following
Clearly, this conclusion does not hold — Daphne may indeed be a cat, but she could just as well be
a woman, a mongoose, or a hippopotamus.
Despite being logically flawed, abduction does seem to play a significant role in language
acquisition, and hence in language change. A relevant example concerns the ‘prepositional object’
of Spanish. Contrast Vi la pared / *Vi a la pared with Vi a la mujer / *Vi la majer. As a rule,
definite human objects in Spanish require the preposition a, even when they are direct objects. This
state of affairs is not found in Latin — how did it arise?
The dative is the rarest case in Latin — almost its only function is to encode participants who are
affected by an action: beneficiaries or ‘maleficiaries’. The Latin case system does not survive into
Spanish, and the functions of the dative are taken over by the preposition a (< AD). (Note that we
must distinguish this dative use of a from the lative [sic] use, encoding motion to or towards: in
Spanish they behave differently with respect to pronominalization: compare Di un libro a Santiago
[person] > Le di un libro with Voy a Santiago [place] > *Le voy.) The overwhelming majority of
participants affected by an action are definite and human — a trawl through texts will confirm this,
as will a moment’s reflection (one generally gives presents, sings songs, reads letters, etc. to an
identifiable human being rather than an inanimate object). Consider, therefore, the following
abductive reasoning:
A child reasoning along these lines will make all definite human objects, regardless of heir function
in the sentence (direct or indirect ibject) daive. In other words, the semantic and pragmatic
characteristics of the object come to determine its grammatical form through a process of abduction.
On abduction in language change, see especially Andersen (1973), and, for a rejoinder, Deutscher
(2002).
Some definitions... A substratum (or substrate) language is one that is already spoken in an area
when speakers of another language arrive and become dominant — socially, culturally, politically,
and linguistically — but which none the less influences this new language. A superstratum (or
superstrate) language is one that is spoken by new arrivals in an area who, however dominant
socially and politically, ultimately fail to dominate linguistically, but still influence the language
which is already spoken in the area. An adstratum (or adstrate) language is one that influences
another ‘from the side’, as it were, without any mass geographical displacement of speakers.
Schematically:
7
© John Charles Smith 2015 Language Change and Historical Linguistics
SUPERSTRATUM
↓
language ‘L’ ← ADSTRATUM
↑
SUBSTRATUM
Substratum hypotheses come with a health warning! Let us take Romance as an example. The
major problem is that the languages invoked as influences were in most cases displaced by Latin
and often had no written form; we therefore have little evidence of them, and cannot assess the
claims with any confidence. Sometimes, the only ‘knowledge’ we have of a substratum language is
its alleged influence on Latin! Beware of this type of circular argument.
FURTHER READING: A sober rehearsal of the problems associated with substratum hypotheses in
Romance is Wanner (1977).
Even superstratum and adstratum languages may present this type of difficulty (although normally
to a lesser degree; but see the comments on Frankish below). The fact that we often have modern
languages that are ‘related to’ (for this metaphor, see §1.7.1) the earlier ones doesn’t always help us
to disentangle the problems: Spanish is ‘related to’ Latin, but is very different from it!
A particularly interesting language from this point of view is French. The major substratum
influence invoked for French is Celtic, specifically the Gaulish language which is known to have
been spoken in Gaul before (and for some time after) the arrival of the Romans. Our knowledge of
this language is limited to several hundred largely fragmentary inscriptions. It is related, at some
distance, to Welsh and Irish, but also has some affinities with the Italic languages, of which Latin is
one. The major superstratum influence is Germanic, specifically the Western Germanic Frankish
language spoken by the Germanic tribe which occupied the northern part of what is now France.
We have even less direct knowledge of this language, although it is assumed to be the ancestor of
modern Dutch. Other parts of Gaul were occupied by tribes speaking Eastern Germanic languages
— the Visigoths (albeit briefly) in the south and the Burgundians in the east. It has been argued
that the tripartite linguistic division of Gaul/France into French (langue d’oïl), Occitan (langue
d’oc) and Franco-Provençal (map on facing page) can be attributed to these different superstratum
influences — see George Jochnowitz (1973), who also points out that the linguistic boundaries
correlate with non-linguistic differences such as traditional legal systems, patterns of crop rotation,
and the pitch of roofs on houses. The major adstratum influences in the history of French have been
Latin (as the language of the church, law, and scholarship), Italian (through cultural and court
influence in the sixteenth century), and, more recently, Arabic (as spoken in France’s North African
colonies) and English.
Different types of ‘colonization’ and ‘invasion’:
The colonizers send an administrative élite, but don’t really populate the new territory
(the Romans in Gaul? — compare the British in India)
The colonizers populate the new territory
(the Romans in Dacia (now Romania)? — compare the British in Australia)
The colonizers settle in the new territory, but adopt its language
(the Franks in Gaul — compare the Normans in England)
8
© John Charles Smith 2015 Language Change and Historical Linguistics
Map of the Gallo-Romance area, showing major language and dialect divisions. From Günter
Holtus, Michael Metzeltin & Christian Schmitt (eds.), Lexikon der romanistischen Linguistik: Band
V, (Tübingen: Niemeyer, 1990), adapted from Pierre Bec, Manuel pratique de philologie romane (2
tomes) (Paris: Picard, 1970-1971).
Note that this neat map, with its well-behaved boundaries between ‘languages’, ‘dialects’, and
‘dialect areas’, represents a massive simplification of the true picture. In fact, there is a multitude
of transitional zones (some of which are represented by the hatching on the map). Isoglosses (lines
which separate areas with different linguistic characteristics) are at best an idealization. See
Chambers & Trudgill (1998), chapter 7: ‘Boundaries’ and chapter 8: ‘Transitions’.
9
© John Charles Smith 2015 Language Change and Historical Linguistics
Why, by the way, did the Franks, who were the socially dominant group and had their own
language, adopt Latin? Latin was the language of an established culture and civilization, which
could certainly be seen as more advanced than that of the Franks; but this sort of argument is
ultimately subjective. The Franks had laws and literature, too. There is, however, one striking
objective cultural difference between the two languages — unlike Frankish, Latin was written.
Compare Toon’s comments on Old English (Toon 1983:2):
‘A preliterate people had conquered a new land and had established there a new social order. [...]
Members of a culture as pervasively literate as our own need to be reminded of the magnitude of that
change. In oral societies, culture and history reside exclusively in memory; social and political structures
depend entirely upon present and immediate balances of power. Spoken agreements are valid only as
long as those who spoke them choose to remember them; material possessions are limited to what an
individual can grasp and defend. Literacy makes possible a past independent of memory. Written
documents outlive convenience and document ownership; bookkeeping makes possible more complex
administrative structures; a king can make a permanent ally by granting a privilege for all time in a
written charter.’
However, this can’t be the whole story, as many Germanic languages that came into contact with
Latin (such as the Old English that Toon is talking about) evolved their own written form as a
result. There must have been other cultural and/or demographic factors at work, as well. For
instance, it’s possible that — like the Normans in England — the Franks, despite their power and
influence, just weren’t numerous enough to impose their language in the long term.
10
© John Charles Smith 2015 Language Change and Historical Linguistics
Latin
Latin fragmented into many different dialects, essentially a continuum; subsequently a handful of
these dialects were ‘selected’ to become ‘standard’ languages.
Latin
(In fact, even this representation is not strictly accurate, because the ‘standard’ language is often
based on a dialect which has been koineized through contact — see §1.9 below.)
Gaulis h
Latin
French
Frankis h
11
© John Charles Smith 2015 Language Change and Historical Linguistics
1.7.3. Are they both right?
Recent work by Labov (2007) seeks to reconcile these two approaches to change in an interesting
way, by claiming that the ‘family tree’ model accounts for linguistic changes which are the result of
child language acquisition, whilst the ‘wave’ model must be invoked to explain change resulting
from imperfect learning by adults. Once again, the L1/L2 distinction turns out to be crucial.
A lot of ink has been spilt on trying to answer the question: Do languages grow more simple or
more complex? Hand in hand with this debate, some scholars have argued that languages evolve in
the direction of greater ‘efficiency’. However, we don’t really know what an efficient language
looks like, and, if we were to try to define the concept, we’d probably begin by saying that an
efficient language was one that stood still — change in itself is communicatively inefficient.
Needless to say, that doesn’t help us! There’s another problem, too — what’s efficient for the
speaker isn’t what’s efficient for the hearer. For the speaker, the ideal language would be some
form of telepathy — expend no physical energy at all and still have your interlocutor understand
every nuance of what you want to say. What hearers would find helpful, on the other hand, is
repetition and redundancy (you’ll observe that I’ve used two words there), so that if they miss
something first time round, they can find it elsewhere in the message. In other words, what
speakers ideally want is incompatible with what hearers ideally want — and, just to make matters
worse, we are all both speakers and hearers, so this battle is constantly being fought out in the mind
of every one of us.
However, the debate on simplification vs. complexification has recently been placed on a firmer
footing by Trudgill (2010; 2011), who begins by pointing out a paradox. Sociolinguists claim that
language contact leads to simplification (often through the process known as koineization or
dialect levelling — see, for instance, Siegel 1985), whereas linguistic typologists see contact as a
cause of complexification. Who is right?
Note, too, that an interesting (and mildly controversial) conclusion to be drawn from Trudgill’s
work is that languages are not equally complex. The ‘equal complexity’ hypothesis has been an
article of faith for many linguists. But, as Ruhlen (1994:148) notes: ‘all extant human languages
are today considered of equal “complexity” by virtually all linguists, despite the fact that there is no
recognized way of measuring complexity in language’. Compare also Lewontin (1990:740): ‘We
would all agree that a human being is more complex than a soap bubble [...] but we cannot all agree
that a dog is more complex than a fish, although fishlike forms preceded doglike forms by 500
million years and were their ancestors. [...] What in fact do we mean with complexity?’
It is clear from the foregoing that language change is not independent of the type of society in
which the language is spoken.
Milroy (1987) investigated the role of social networks in language variation and change, studying a
number of communities in Belfast, the capital of Northern Ireland. Speakers will characteristically
have interchanges with: i) members of their family; ii) their friends; iii) their neighbours, and iv)
their workmates. The greater the overlap between these categories of people (i.e., the denser the
network), the higher the network strength score of the individual involved. The broad finding
from Milroy’s work is that the greater an individual’s social network strength score, the more likely
that person is to exhibit local (i.e., non-standard) variants. This has a number of important
consequences. One of these is that the role of variables such as social class, gender/sex, etc. in
language variation and change may be epiphenomenal. In some of the areas studied by Milroy,
men went out to work, whilst women did not; moreover, when they married, women tended to
move to live with their husbands rather than vice versa. Women therefore had a lower network
strength score on two counts (no interchanges with workmates; fewer interchanges with family). In
these communities, Milroy observed the ‘expected’ pattern of sociolinguistic variation — men used
more local variants and women more standard variants. However, in one area, there was high male
unemployment and most women went out to work (in the same place). Women therefore had a
higher network strength score than men. Here, the ‘expected’ pattern was not observed — rather
the reverse. Women used more local non-standard variants than men. It appears therefore that
linguistic behaviour, at least in the communities studied by Milroy, is not directly correlated with
variables such as social class, gender/sex, etc., but rather with network strength score.
14
© John Charles Smith 2015 Language Change and Historical Linguistics
1.10.2 Societal typology and language change
Andersen (1988) and Røyneland (2004), Kerswill (2015), building on the work of Milroy and
others, distinguish two typological parameters of societal variation: a sociocommunicational or
demographic parameter (open vs. closed) and an attitudinal or ideological parameter
(endocentric vs. exocentric).
Closed societies are essentially societies in which the majority of the population have high network
strength scores (à la Milroy).
Open societies are essentially societies in which the majority of the population have low network
strength scores (à la Milroy).
Endocentric societies set great store by solidarity within the community and are suspicious of or
hostile to outside influence.
Exocentric societies accept or even welcome outside influence and regard solidarity within the
community as secondary.
This yields a four-way typology, as follows.
Closed endocentric societies. Characteristically, these are isolated rural communities, although
many metropolitan inner cities now also conform to this definition. Inner London, where a strong
sense of community holds amongst people of many different ethnic and geographical origins has
led to a high degree of linguistic contact and the emergence of ‘new London English’ or
‘multicultural London English’ (see Cheshire et al. 2011, Kerswill 2015).
Open endocentric societies. These are typically urban centres with both a strong sense of
community and strong external contacts. They tend to be sources of innovation, from which new
forms diffuse outwards; they are ‘optimal senders’ of language change. Cities such as London and
Paris have played, and to some extent continue to play, this role, which is also played at a regional
level in Britain by Manchester, Liverpool, and Glasgow.
Closed exocentric societies. These are low-contact communities which none the less have a
positive orientation to outside norms. Because of this orientation, they are receptive to change, but,
in Kerswill’s phrase (Kerswill 2015), this is essentially ‘change driven by ideology, not change
driven by contact’. A good example in the U.K. is inner-city Glasgow.
Open exocentric societies. These are typically very small communities, which are being subsumed
into another community’s identity. They are therefore likely to be transitional and unstable. These
are often rural communities (compare Marshall 2004 on the village of Huntly in Aberdeenshire in
Scotland), but may also be mobile suburban areas (such as outer-city Belfast). They are ‘optimal
receivers’ of language change.
What emerges with some force from these recent studies, is that endocentricity and exocentricity
(i.e., attitudes) appear to be more significant determinants of language than closed vs. open
networks. Note that there had been indications that this may be the case since Labov (1963).
15
© John Charles Smith 2015 Language Change and Historical Linguistics
fashionable for British rock musicians to attempt (and often fail) to imitate an American accent.
With the success of British rock music, the reverse became fashionable. It was at about this time
that ‘Punk Rock’ bands, made up of young working-class south Londoners came to the fore. Many
of their speech habits began to propagate through the community as a result of ‘covert prestige’ —
very schematically, they came to mark groups of adolescents and young people who wished to
distinguish themselves from (or even rebel against) the generation of their parents. Subsequently,
they came to mark an identity as a young person in a broader sense.
That seems to be that — a specific, circumscribed change, which has taken place in the last thirty or
forty years, and which has a specific, almost punctual, social motivation. But if, instead of looking
at this change through a microscope, we shift focus and use a telescope, an interestingly different
picture emerges. At some stage in the mid- to late seventeenth century — certainly by 1700 —
non-prevocalic [r] started to disappear from British English. This was a gradual development; it
appears to have started in London and spread slowly through the country. There are pockets of
resistance to this day — south-west England, parts of Lancashire, and parts of Northumberland are
still rhotic, as, of course, are most of Scotland and Ireland. The change did not affect much of
North America, which was separated from England by a vast distance (and much of which would
shortly become politically separate, as well), except in a handful of words (cuss < curse, bust <
burst, etc.) — thereby indicating that the change was probably lexically diffuse. But the change
spread out from London and gradually reached most of England. It appears to have begun as a
vocalization of the [r] to schwa; thereafter, in some postvocalic contexts, the schwa fell, leaving us
with the situation we have today.
Now, [l] and [r] are members of the same small class of sounds — the liquids. Indeed, in English,
they are the only two members of this class. A linguistic historian of the year 3000 is going to tell
the following story: there was, in the history of English at the end of the second millennium, a
change which we might refer to as ‘liquid vocalization’. First of all, it affects [r], beginning in the
late seventeenth century. Only about three hundred years later is [l] affected; but we are clearly
dealing with a single, if somewhat drawn out, change.
The microscope offers us a detailed sociolinguistic account. The telescope gives us a structural and
systemic account. The unanswered question is: how do we reconcile the two?
‘When did Latin become Spanish, French, Italian, etc.?’ and ‘Why didn’t Chinese stop being
Chinese?’ are diachronic analogues of the famous (non-)question: ‘What is the difference between a
language and a dialect?’. The answer is the same: two speech varieties are different languages
(synchronically or diachronically) if that is how speakers perceive them (or wish to perceive
them). The question is cultural and ideological, rather than linguistic stricto sensu.
16
© John Charles Smith 2015 Language Change and Historical Linguistics
This issue is discussed, with specific reference to Romance, by Lloyd (1991) and Janson (1991).
Janson quotes the example of a group of southern African language varieties, which were mutually
intelligible and had no separate names. However, different groups of their speakers were
proselytized by missionaries from Britain, who named the language Setswana, missionaries from
France, who called it Sesotho, and missionaries from Germany, who designated it Sepedi. There
thus grew up a new sense of linguistic separateness which had not existed previously, even though
nothing about the language itself had actually changed. Similar examples abound. Chambers &
Trudgill (1998:9-12) point out that the area known as Skåne, now the southernmost region of
Sweden, was Danish until 1658, when it was conquered by the Swedes. Until then, the inhabitants
had claimed to be speaking a dialect of Danish; a generation later, with their language practically
unchanged, they claimed to speak a variety of Swedish. The language had not changed, but social,
political, and geographical perceptions of it had (to use the appropriate technical term, it had been
heteronomous with — i.e., perceived as a dialect of — Danish, but became heteronomous with
Swedish). Equally striking is the example of Afrikaans (also mentioned by Chambers & Trudgill,
loc. cit.); until the 1920s, this variety was considered heteronomous with Dutch, but, with the
advent of greater Afrikaner nationalism, both Dutch and South Africans came to view it as an
autonomous language. A slightly different, but related, phenomenon is discussed by Gerritsen
(1999).
Crucial in this context are the Carolingian reforms of c. 800 AD. Charlemagne (c. 742-814) was
King of the Franks from 768 until his death and Emperor from 800. He was not a native speaker of
Latin/Romance, and surrounded himself with courtiers who were likewise non-native speakers (one
of the most famous of whom was Alcuin of York, c. 735-804). By this stage, the spoken vernacular
had developed considerably since the Classical Latin period (the first centuries BC and AD), but the
written language had remained more or less the same. There seems to have been a feeling that the
gulf between the written and spoken language should not be so great, perhaps coupled with a certain
nostalgia for the days of great authors such as Cæsar, Cicero, and Livy. So, whereas the obvious
resolution of the discrepancy might have been to bring writing closer to speech, the opposite path
was taken, and it was determined that the language should be pronounced as it was written. It’s as
if a new ruling class, who were not native speakers of English, arrived in twenty-first-century
Britain, observed the huge differences between the language’s phonology and its spelling system,
and decreed that, since the spelling represented the way English was spoken in a heyday of its
literature (which it does — it’s a pretty accurate reflection of the pronunciation that Chaucer would
have been used to), it should henceforth be given a literal realization. Although only a small
minority of the population could read and write during the Carolingian period, the reforms were far-
reaching. Church services, which most people attended, suddenly became incomprehensible
because of the new pronunciation; and, whilst the ritual parts of the Mass could be understood
precisely because of their ritual nature, sermons fared less well. The Council of Tours, in 813,
sought to remedy this problem by stipulating that priests should preach in rustica romana lingua —
i.e., the way ordinary people spoke, rather than a letter-by-letter pronunciation of the written
language. It is obvious that these circumstances created for the first time a situation which favoured
a conceptual distinction between ‘Latin’ and ‘the vernacular’ (although it would be several
centuries before this vernacular was given a precise name in any Romance-speaking area). An
associated issue is that it soon became necessary to write the vernacular down; and, since it wasn’t
Latin any more, a new spelling system had to be developed, by trial and error. We shall have cause
to return to these points. They are addressed most famously by Wright (1982), and, latterly, from a
different perspective, by Banniard (2013).
17
© John Charles Smith 2015 Language Change and Historical Linguistics
3. Methodological Preliminaries
All diachronic linguistics and the synchronic study of any language other than in its contemporary
state (even, say, something as recent as nineteenth-century English or Spanish) involves using ‘dead
data’. This fact constrains us in significant ways.
3.1. Sources denied to us
Until very recently, speech in general. We can never be certain, therefore, how a ‘dead’ language
was pronounced; this fact, in turn, can vitiate claims about its phonetics and its phonology.
By definition, native-speaker judgements and intuitions. This limits us, by depriving us of a
natural mode of investigation, and raises the fundamental issue of attestation (see below).
3.2. Sources we do have at our disposal
Essentially, written texts — a ‘dead’ language is inevitably a corpus-based language or ‘text
language’. (Even sound recordings of earlier stages of a language, whether wax cylinders or .mp3
files, constitute a finite corpus.) A corpus-based approach to any language can yield useful results.
However, some scholars have objected to the use of corpora in linguistic research. They argue that
the aim should be to investigate the system that underlies, or lies behind, the data (in other words, to
study competence rather than performance), and that a linguistic hypothesis must apply to the
potentially infinite set of sentences (etc.) of the language, not just to some subset of utterances
(etc.). This view is now rather less prevalent than it used to be. And, anyway, in the case of a
‘dead’ language, the subset (and often a rather random subset at that) is all we’ve got — if it is to
take place at all, the study of a ‘dead’ language has to adopt a corpus-based approach.
3.3. The problem of attestation and the issue of ‘negative evidence’
If a form or structure is not present in our corpus, does this mean it is absent from the
language, too; or is its absence accidental? In the case of a living language, we would go and
check with a native speaker — but, of course, with a ‘dead’ language, we can’t. We can make
intelligent guesses — but, methodologically, that is all they’ll ever be... The importance of the
‘uniformitarian hypothesis’ (see §1.5.2 above) may become clearer at this point.
3.4. The unrepresentative nature of the corpus
Most of what has survived is the result of cultural transmission — that is, texts with a
perceived literary or historical value. Such texts tend to be formal, conservative, or even archaic.
At the other extreme, they may be experimental! Compare George Steiner’s definition (Steiner
1971) of literature as ‘language in a condition of special use’. Ephemera (almost by definition!)
tend not to survive, or not to anything like the same extent. (There are, of course, famous
exceptions, such as the Pompeian graffiti. Some commercial and legal documents also survive
from the Middle Ages.) In addition, there is an element of randomness about the corpus, due
sometimes to human factors and sometimes to natural disasters — manuscripts have been lost or
destroyed in fires or floods; conversely, the inscriptions in everyday Latin scratched on walls at
Pompeii and Herculaneum have come down to us only because they were buried by the eruption of
Vesuvius in 79 A.D.
All of the foregoing means that our perception of the language is skewed — imagine a description
of contemporary English which was based almost entirely on poems, novels, and (to a lesser
degree) plays, together with a handful of wills and mortgage deeds, a shopping-list or two, and the
odd spray-paint ‘tagging’ of a railway carriage or motorway bridge, and which took no account at
all of the spoken language or of other written registers, such as journalism.
18
© John Charles Smith 2015 Language Change and Historical Linguistics
A majority of mediæval literature is poetry — literary prose is a relative newcomer. This clearly
causes problems, because poetic form imposes special constraints (metre, alliteration, assonance,
rhyme, etc.). However, paradoxically, these same poetic conventions can sometimes help us in
determining how something was pronounced (or even whether or not it was pronounced). For
instance, in Latin verse, when two vowels came into hiatus at a word boundary (i.e., when one word
ended in a vowel and the next word began with a vowel), this yielded a single (vowel for purposes
of scansion. Now, when the first word ended in ‘vowel + <M>’ and the second word began with a
vowel, exactly the same thing happened — the sequence ‘vowel + <M> + vowel’ was also counted
as a single vowel. Thus, schematizing, both BONA ET VERA AMICA and BONAM ET VERAM AMICAM
‘good and true (female) friend’, in the nominative and accusative, respectively, would both be
scanned as BON’ ET VER’ AMICA(M). Poetry (certainly Latin poetry) is usually formal and
conservative. The obvious conclusion to be drawn is that final [m] was no longer pronounced in
any register; and, indeed, we have evidence from other sources for this development. Likewise, we
find that, in the earliest assonating Old French verse, a vowel followed by a nasal consonant may
assonate with the same vowel in any other environment. When this ceases to be the case, and we no
longer find vowels followed by nasal consonants in assonance with vowels not followed by nasal
consonants, we may tentatively assume (despite the fact that the evidence is negative rather than
positive — see §3.3 above) that phonemic nasal vowels have emerged, distinct from their oral
counterparts.
19
© John Charles Smith 2015 Language Change and Historical Linguistics
4. A case-study in language-contact: Latin and Romance influences on English
4.1. Typology of contact (see §1.6.2 above)
4.1.1. Substratum
4.1.1.1. Latin in Britain
Latin was spoken in Britain before the Germanic invasions of the fifth century. It would therefore
have been possible for Latin words to have entered Old English from this source. There appear to
be few, if any, examples of this happening. One candidate is the word street, from Latin STRATA,
the feminine past participle of STERNERE ‘to lay out, pave’. Some other Romance words, such as
cross, did enter English during this period, but as adstratum elements (see §4.1.3. below).
4.1.1.2. Spanish in the United States
Many centuries later, Spanish was arguably a substratum of U.S. English; it was a language spoken
in Florida and the south-western states before they were conquered by ‘Anglos’, and some words
(pueblo, adobe) entered American English as a result. In practice, however, it is difficult to
disentangle substratum and adstratum elements here.
4.1.2. Superstratum
England was conquered by the French-speaking Normans in 1066, and Norman French (or ‘Anglo-
Norman’) was the dominant official language of the country until the late fourteenth century.
Although the ruling class eventually adopted English as their language, it was an English which had
absorbed many Norman French words. ‘Many of the words borrowed [sic] at this stage have
become so much part of the language that many modern speakers find it difficult to think of them as
loan [sic] words’ (Blake 1992:17).
4.1.3. Adstratum
Latin was the language of the church and of scholarly endeavour throughout the Middle Ages.
French remained a fashionable language in high society, and, from the eighteenth to the twentieth
century, was the major language of international diplomacy. Italian and Spanish also had a
cultural cachet. Spanish and Portuguese (and, to a lesser extent, French) existed alongside
English during the colonization of the New World. English absorbed words from all these sources.
Note that different varieties of English sometimes have different adstrata. Compare, for instance:
For obvious reasons, Spanish has been a much more significant adstratum for U.S. English than it
has for British English.
4.2. Semantic typology of Latin and Romance elements in English
4.2.1. Superstratum items (Norman French)
According to Burnley (1992:429), ‘The very earliest loan [sic] words from French appear in pre-
Conquest documents, and reflect aristocratic values and tastes’. (They might therefore in fact
qualify as adstratum elements.) The social dimension of superstratum in Middle English is shown
by the well-known contrasts between the names for animals and the names for their meat: ox, sheep,
pig vs. beef, mutton, pork. The English-speaking peasant reared the living animal; the French-
speaking aristocrat was more likely to encounter it on a plate.
20
© John Charles Smith 2015 Language Change and Historical Linguistics
4.2.2. Adstratum items
These words are often broadly culture-bound; that is, they are (for obvious reasons) connected
with the fields in which the languages were pre-eminently used.
Latin church requiem, magnificat, etc.
law executor, arbitrator, proviso, alias, alibi, etc.
academia alma mater, viva (voce), etc.
French diplomacy chargé d’affaires, attaché, coup d’état, etc.
food gourmet, restaurant, café, menu, etc.
fashion haute couture, prêt à porter, etc.
Italian music aria, sonata, toccata, (violon)cello, etc.
Note that Italian words such as cappuccino, pasta, spaghetti are narrowly rather than broadly
culture-bound, in that they relate to specifically Italian items. Likewise, Spanish adstratum items
tend to be more narrowly culture-bound, relating chiefly to Spanish (and Latin-American) food
(e.g., tapas, gazpacho, tacos) and activities such as bull-fighting (toreador, matador, etc.);
however, note aficionado and incommunicado (Spanish incomunicado, with the <m> doubled on
the analogy of communication, etc.). Ironically, the word auto-da-fé, referring to a sentence of the
Spanish Inquisition and its execution, comes from Portuguese!
Other Romance adstratum sources are rare, and tend to be restricted to words denoting narrowly
culture-bound concepts which are difficult to translate, or where use of the original language adds
‘local colour’ to a text, e.g., Romanian securitate ‘[Ceauşescu’s] secret police’; Catalan generalitat
‘government of Catalonia’. Perhaps the best example is the Rumantsch language (spoken in eastern
Switzerland), whose sole contribution to English is its own name!
A semantic typology of French words in English (adapted from Chirol 1973)
Food
Fashion Manufacturing Terms
Types of garment
Materials
Hairdressing and Cosmetics
The Home Furniture
Architectural Terms
Motor Cars
Leisure Games and Sports
Parties, Dancing, Theatre
Travel and Tourism
Culture Literature
Art Painting
Sculpture
Architecture
Music
Society Organization
Behaviour
Relations Politeness
Conversation
Love
‘La masse des gallicismes acquiert ainsi une singulière cohérence ; ce que l’anglais emprunte [sic] à
la France à travers ses mots c’est un art-de-vivre et un savoir-vivre.’ [‘In this way, the great
21
© John Charles Smith 2015 Language Change and Historical Linguistics
majority of gallicisms form a strikingly coherent group; what English borrows from France through
its vocabulary is an art-de-vivre and a savoir-vivre.’] (Chirol 1973:32)
‘Déplacé du sémantique au stylistique, du dénotatif au connotatif, de l’objectif au subjectif, le statut
du mot devient entièrement verbal. [...] Il est clair qu’aujourd’hui — et sans pour autant nier
l’importance des relations et des échanges techniques, scientifiques et culturels — la majeure partie
des gallicismes de l’anglo-américain assument cette fonction purement mythique.’ [Through a shift
from semantic values to stylistic values, from denotation to connotation, from objectivity to
subjectivity, the word ends up being intrinsically just a word. [...] It is clear that today, for all the
importance of contacts and of technical, scientific, and cultural exchange, the majority of French
elements in British and American English have the function of creating or perpetuating a myth.]
(ibid:30)
However, there is a tremendous number of French words which have made their way into English
since the Renaissance without belonging to any identifiably culture-bound semantic field —
compare reservoir, cul-de-sac, secateurs
4.3. Frequency and distribution — types, tokens, and register (see §1.6.1.2 above)
4.3.1. Frequency
4.3.1.1. Middle English
Baugh (1935) notes that just over 10,000 French words were adopted by Middle English, of which
about 75% remain in use. Barber (1976:166-168) takes a 2% sample of the OED. 697 words are
recorded as arising in Middle English (derivations as well as foreign words). The number of these
items taken from French is 223, or 32%; the number from Latin is 59 (8.46%), and the number that
are ambiguous as between French and Latin 12 (1.72%). The number of new items entering
English from French peaks in the late fourteenth century (interestingly, at a period when French as a
means of communication in England had long been in decline). According to Dekeyser (1986),
91.5% of the Early Middle English lexicon consisted of etymologically ‘English’ words; by the
Later Middle English period, the figure was 78.8%. However, an analysis of tokens (occurrences)
rather than types (dictionary entries) (see the discussion in §1.6.1.2) yields figures of 94.4% and
87.5%, respectively. Although many French words had entered the language, they were not used as
frequently as ‘native’ words. In particular, relatively few functional words (prepositions,
complementizers, etc.), which are characteristically of high frequency, are of French origin (but see
§4.5.1.1. below).
4.3.1.2. Modern English
Barber’s survey contains 1,848 words which enter English between 1500 and 1700. 625 of these
are taken from other languages. 121 are from French (19.4% of foreign words; 6.55% of all new
words), 393 from Latin (62.9%; 21.3%), 16 from Italian (2.56%; 0.866%) and 16 from Spanish or
Portuguese (2.56%; 0.866%). 20 words (3.20%; 1.08%) are ambiguous as between a French origin
and a Latin one. The total number of Latin/Romance words is 566 (90.6%; 30.6%).
Estimates of the number of Latin/Romance words in contemporary English vary; but, here, too, the
proportion of types is undoubtedly greater than the proportion of tokens. (‘Although the majority
of the words contained in modern English dictionaries are, immediately or ultimately, of Latin
origin, English is nevertheless, by [...] its everyday vocabulary, a Teutonic [sic] language’ (Weekley
1924:ix).) Of the original ‘Thorndike 500’ (the 500 commonest word-tokens of English based on a
broad corpus of material, but selected with the aim of teaching children to read — see Thorndike &
Lorge 1944), 79 (or 15.8%) are of Romance (including Latin) origin, as follows:
add, ball, beautiful*, because*, blue, box¶, case, cause, certain, change, city, clear, close, colour,
company, country, course, cover, cross¶, dress, during, face, family, fine, flower, form, fresh,
front, general, hour, just, large, letter, line†, measure, mile, money, mountain, move, paper‡,
part, pass, pay, people, person, picture, piece, place, plain, plant¶, please, point, poor, power,
22
© John Charles Smith 2015 Language Change and Historical Linguistics
present, reason, receive, remain, remember, rest, rich†, river, rock, save, school‡¶, second, serve,
several, sound¶, state, street¶, sure, table, train, turn, use, very, visit, voice
More detailed conclusions can be drawn from more recent (computerized) word counts. Compare
the Romance items amongst the 500 most frequent words in Caroll et al. (1971) (a 5,088,721-token
cor-pus, containing 86,741 types, and comprising 500-word samples from 1,045 texts used in
schools):
people, use, very, just / used, because*, part, place, different, number, air, line†, (set), sound,
large / school‡¶, important, form, animals, page, parts, country, picture, study, second, story,
paper, sentence, during*, sure, (means), miles, example, several, change, turned, point, city,
using, usually / money, car, family, turn, move, face, group, sentences, plants¶, United States
[sic], order, lines†, really*, table, remember, course, front, space, close, idea‡, add, places, letter,
letters, able, (mean), (rest), certain / special, complete, person, state, (list), notice, voice,
probably, area, Indians, sounds, matter, box¶, class, piece, surface, river, numbers common, fine,
round, past, ball, questions, blue, finally*, animal, power, problem‡, Indian, mountains,
beautiful*, moved, system‡
(total (disregarding bracketed homonyms) = 97 or 19.4%: 4% in first 100, 10% in second 100, 24%
in third 100, 26% in fourth 100, 33% in fifth 100; first Latin/Romance word ranks no. 79)
Hofland & Johansson (1982) (Lancaster-Oslo-Bergen corpus of British English: approximately
1,000,000 tokens, consisting of 500 text samples of about 2,000 words each, all published in the
year 1961):
very, people / just, because*, used, course, part, per, fact, case, use, during*, government, place,
quite, number, sir / present, general, large, possible, point, country, face, perhaps*, school‡¶,
local, party, war, second, form, (set), important, different, national, public, (miss), council,
certain, order, round, turned, power, money, (means), really*, group, moment, able, service,
view / matter, necessary, age, system‡, air, common, interest, area, family, probably, table,
labour, reason, question, car, state, voice, effect, social, position, change, sense, company,
members, office, particular, act, problem‡, education, example, period, experience, special,
modern, usually / value, political‡, cent, certainly, particularly, terms, idea‡, turn, (mean),
committee, result, minister, line†, past, art, due, added*, difficult, century, countries, real,
society, concerned, problems‡, doubt, music, stage, human, cases, sure, increase, president, rate,
level, close, clear, except, several, story, results, development, market¶, city, control, type‡,
policy‡, united, various, pay, total, industry, future
(total (disregarding bracketed homonyms) = 134 or 26.8%: 2% in first 100, 15% in second 100,
31% in third 100, 35% in fourth 100, 51% in fifth 100; first Latin/Romance word ranks no. 81)
and Francis & Kučera (1982) (Brown corpus of American English: approximately 1,000,000
tokens, consisting of 500 text samples of about 2,000 words each, from 15 different text types, and
all published in the year 1961):
state, use [v.], people, Mister, just / because*, very, school‡¶, number, part, during*, place [n.],
turn, area, problem‡, system‡, long, Mrs, fact, course, city, group, case, government, unite,
point, line†, country, provide, member / company, service, move, form, president, order, interest,
family, power, car, figure, use [n.], value, social, face, several, question, general, development,
possible, national, force, per, important, action, large, nation, appear, student, doctor, cost,
example, matter, continue, reason, require, idea‡, expect, second, increase / result, around*
[prep.], hour, pay, street¶, develop, consider, remain, type‡, sense, different, period‡, college,
perhaps*, effect, experience, public, study, rate, office, moment, serve, person, pass, history,
receive, class, add, position, policy‡, court, change [n.], method‡, party, society, local, plan,
section, age, money, community, determine, effort, department, condition, information, material,
level, voice, probably*, human, art, centre, air / letter, include, political‡, university, century,
process, remember, involve, situation, federal, available, special, indicate, economic‡, minute,
23
© John Charles Smith 2015 Language Change and Historical Linguistics
table, real, tax, around* [adv.], purpose, act, term, place [v.], Miss, return [v.], activity,
difference, picture, really*, spirit, change [v.], sure, present, major, stage, necessary, pressure,
property, industry, certain, surface, control, piece, able, offer, view, (arm), produce, music‡,
education, record, enter, union, story, common, apply, secretary
(total (disregarding bracketed homonyms) = 180 or 36.0%: 5% in first 100, 25% in second 100,
40% in third 100, 54% in fourth 100, 56% in fifth 100; first Latin/Romance word ranks no. 64)
4.3.2. Register In general, more ‘elevated’ texts tend to contain more words of Latin/Romance
origin.
Driving School, BBC1. Joan Rodwell was desperate to learn to drive so she could take her
grandchildren out. She’s had 110 lessons with five driving instructors and her partner Brian
was so helpful that when they went out in the car he made her sit in the back while their Old
English Sheepdog Jake sat in the front. When Joan took her test, she panicked so badly she
couldn’t even move the car for ten minutes — but she passed. On that form I think the
examiner would have passed Stevie Wonder. Joan has a wonderful winning smile but I think
I’d rather be driven by Jake. This hilarious fly-on-the-windscreen look at the frightening
business of learning to drive was highly entertaining — and it should do wonders for public
transport.
(The Sun, 11 June 1997)
(tokens: 22/132 = 16.7%; types: 20/89 = 22.5%)
Another example of lite-tv that got under way last night was Driving School (BBC1) and very
watchable it was, too. It was watchable (as lite-tv is designed to be) but so contrived you
wondered just how spontaneous some events actually were. Had Maureen (six test failures,
400 lessons) really steered her husband’s elderly Lada into the path of an overtaking car? Had
she really woken him up at 2 o’clock in the morning with the words: ‘It’s no good, love, we’ve
got to do it’? Revise for her theory test, that is. In her efforts not to miss a single moment that
could be made humorous, the director, Francesca Joseph, had deployed cameras everywhere.
In the cars, in the bedroom, in the underwear department of a well known chain store...
(The Times, 11 June 1997)
(tokens: 33/132 = 25%; types: 28/97 = 28.9%)
24
© John Charles Smith 2015 Language Change and Historical Linguistics
Note: BBC, although standing for ‘British Broadcasting Corporation’, is treated as unanalysable.
However, ‘tv’ is assumed to represent ‘television’. The origins of proper names are ignored.
Contracted forms and numerals are counted as one word, compounds as multiple words.
25
© John Charles Smith 2015 Language Change and Historical Linguistics
'vəʊsɪ] or ['vaɪvə 'vəʊtʃe], perhaps influenced by Italian sotto voce (> English ['sɒtəәʊ 'vəʊtʃe] )), and
words entered the language in this form. Romance words posed greater problems.
‘In M[iddle] E[nglish] times it was common for loanwords [sic] from French and Latin to spread
from upper-class use to lower-class speech, becoming variously adapted in the process. Such
integration was much less effective with loans [sic] from contemporary languages from the
sixteenth century on. Many words, in particular those derived from French and Italian, appear to
have been left unadapted because they were regarded as shibboleths of educated speech.’ (Görlach
1991:156-157)
4.4.2.1. Incorporation into system
Segmental adjustments:
accommodation diphthongization ([e] > [eɪ]; [o] > [əʊ], etc.)
unrounding ([ø] & [œ] > [ə:]; [ɥ] > [w])
reduction of unstressed vowels ([a] > [ə], [i] > [ɪ], etc.)
miscellaneous ([a] > [æ]; [ʁ] > [ɹ]; [t̪] > [t]; [p-] > [ph-], etc.)
linearization (e.g., French [y] > [ju]; [ɑ̃] & [ɔ]̃ > [ɒŋ]; [ɲ] > [nj] or [jn])
redistribution (e.g., deletion of postvocalic /r/)
Suprasegmental adjustments:
stress reassignment, resyllabification
26
© John Charles Smith 2015 Language Change and Historical Linguistics
4.4.2.1.1.4. Stress patterns of disyllabic French adstratum nouns
In U.K., often accommodated to the trochaic stress pattern normal for English nouns.
In U.S.A., often given iambic stress, presumably to match the final phrase stress of French.
27
© John Charles Smith 2015 Language Change and Historical Linguistics
during — by grammaticalization of a present participle) are a locus for incorporation, betokening
the profound importance of the superstratum language.
4.5.1.2. Morphological adjustments
4.5.1.2.1. Roots
As expected, superstratum items are generally accommodated to an unmarked morphological
class. Thus, most imported nouns form regular plurals in -s, and most verbs enter the weak
conjugation, forming both their past tense and past participle by suffixing -ed. However, there are
rare, but striking, examples of verbs being accommodated to a strong conjugation — witness catch
(< cachier; compare Modern French chasser ‘to hunt’), caught, caught (allegedly by analogy with
the semantically similar ‘native’ verb latch ‘seize, grasp’; and strive (< estriver), strove, striven (by
analogy with drive), although strived is also found as both past tense and past participle.
4.5.1.2.2. Affixes
Several affixes which originally occurred only in superstratum items came to be used with ‘native’
roots. Examples are the nominalizing suffixes -age (compare Modern English luggage, roughage)
and -ment (compare Modern English betterment, fulfilment), the feminine suffix -ess (compare
Modern English shepherdess (hirdess is found as early as Chaucer)), and the adjectival suffix -able
(which occurs only with Romance roots in Middle English, but which is fully productive in the
modern language — compare watchable in line 2 of the second text of §4.3.2.2. above, bearable,
bookable, likeable, unputdownable, etc.). Miller (1997) gives figures for such ‘hybrid derivatives’:
in 1400, he finds 64 of them, in 1450, 100, and in 1500, 130.
4.5.2. Adstratum items
4.5.2.1. Categorial identity
Categorially, adstratum items are more restricted than superstratum items, indicating less
‘intense’ contact — most of them are nouns or noun phrases (although adjectives, such as chic,
have also entered English from this source).
28
© John Charles Smith 2015 Language Change and Historical Linguistics
(Guardian (London), G2 supplement, 7 June 2002, p. 2), instead of grandes dames; Schumacher
won a lot of grand prix instead of ...grands prix) and grammatical gender. In particular, le is often
assumed to be the sole form of the French definite article; thus (from Sheffield, England) le
Boutique as the name of a (particular) fashion shop (cf. French la boutique), and even le Pizza as
the name of a pizzeria, which is impossible in either French (la pizza) or Italian (la pizza, plural le
pizze). Oxford boasts the equally impossible sandwich bar le Panini (Italian i panini ‘the
sandwiches’ (masculine plural); le is masculine singular in French and feminine plural in Italian!).
Recently noted Australian examples include le Maison (cf. French la maison) (Nicholson Street,
North Carlton, Victoria) and le Crepe (cf. French la crêpe) (Bondi Beach, Sydney, NSW). Indeed,
in some registers of English (particularly in the language of advertising), le has come to be used
virtually as a prestige form of the definite article — compare le bag, le car (used to advertise the
Renault 4 — car is clearly an English word here, as in French, this expression would mean ‘the
motorcoach’!), and the train that carries cars through the Channel Tunnel, le Shuttle. This
admittedly highly marginal use of le constitutes an exception to the general principle that adstratum
is not a source of functional items. Compare also ‘Posh Spice Victoria Adams yesterday defended
her vilified soccer superstar fiancee David Beckham’ (The West Australian (Perth, W.A.), 23 July
1998, p. 7) and other examples of her fiancée (cf. Independent on Sunday (London), 5 March 2000).
Fiancée is feminine; the corresponding masculine form is fiancé. But, because English doesn’t
usually mark gender in this way, the form in -ée (often written without the accent) is sometimes
(mis)interpreted as an epicene (i.e., ‘unisex’) patient suffix, perhaps by analogy with divorcee,
employee, internee, refugee, etc.
4.6. Calques
‘Loan [sic] translations’, better known as ‘calques’ (itself a French word meaning ‘tracings’, which
was adopted by English in the 1930s), involve using a Latin or Romance ‘template’, but replacing
the original Latin or Romance elements with the corresponding English items.
29
© John Charles Smith 2015 Language Change and Historical Linguistics
4.6.2. Syntactic calques
Syntactic calques are found, too. For instance, the normal position for an adjective in English is
before the noun it qualifies; but in some contexts, the unmarked French order of these items, in
which the adjective follows the noun, may be calqued. Often, the French word order is found in
terms relating to diplomacy, a field in which French was for long the dominant world language:
thus, the General Secretary of a trade union, but the Secretary General of the United Nations, an
Ambassador Extraordinary and Plenipotentiary, etc. Examples can also be found in other French-
dominated areas, such as heraldry, chivalry, and the law, often leading to doublets of technical vs.
non-technical phrases — thus, bend sinister ‘a diagonal line running from bottom left to top right of
the shield’ (vs. sinister bend ‘a curve which bodes ill’), knight errant ‘a knight travelling in quest of
adventure’ (vs. errant knight ‘a knight of questionable standards or behaviour’), fee simple
‘complete and absolute ownership’ (vs. simple fee ‘a one-off payment which is straightforward and
uncomplicated’), etc.
More widespread, and of greater significance, is the non-finite complementation construction
known as the ‘accusative and infinitive’. Latin verbs of saying, believing, etc., took non-finite
complements formed with an accusative objective and the infinitive of the subordinate verb:
compare CREDO EAM DIVITEM ESSE, lit ‘I-believe her (acc.) rich (acc.) to-be’. The ‘native’ English
construction in these circumstances is one of finite complementation, with or without the overt
complementizer that: I believe that she's rich, I believe she's rich. However, in more formal
registers, the accusative and infinitive is found, apparently calqued on the Latin: I believe her to be
rich, etc. On the plausibility of Latin influence here, see Nagucka (1985) for Old English and and
Fischer (1989) for a historical survey.
4.7.2. Morphology
Forms which are neither the original Latin or Romance nor accommodated to English can arise
when morphology is reanalysed (or ‘misunderstood’). For instance, the Latin second-declension
plural pattern (singular -us, plural -i) is often over-generalized to words like syllabus, prospectus
(*syllabi, *prospecti for syllabuses, prospectuses: Latin plurals syllabus, prospectus), platypus
(*platypi for platypuses: not even a Latin word, but coined (in 1799) from Greek πλατύς ‘broad’
and πούς ‘foot’), and octopus (*octopi for octopuses: likewise Greek, from ὀκτώ ‘eight’ and πούς
‘foot’).
Further reading: Durkin (2014) is a comprehensive recent survey of loan words in English.
Rodríguez González (1995, 2001), Cannon (1996), Gooch (1996), Lodares (1996), and other papers
in the same volume deal with Spanish elements borrowed into English.
30
© John Charles Smith 2015 Language Change and Historical Linguistics
31
© John Charles Smith 2015 Language Change and Historical Linguistics
However, the claim that Latin is synthetic whilst Romance is analytic is much too simplistic. Note
the verb forms in the above sentences. Four of them are synthetic in Latin (EST, MOMORDIT, DEDIT)
and remain synthetic in Spanish (es, mordió, dio); INTERFICIETVR is the only synthetic form of Latin
which corresponds to an analytic form in Spanish (será matado). Likewise, if we look at plurals
and feminines in Latin (CANIS ‘dog’ vs. CANES ‘dogs’; BONVS ‘good’ masc. vs. BONA ‘good’ fem.)
and Spanish (perro vs. perros; bueno vs. buena), we find that the categories of number and gender
receive synthetic exponence in both languages. Clearly, the bald statement that Latin is synthetic
and Spanish analytic is an over-generalization. But is there any sense in which it is true? In fact,
there may be. But before we can investigate this possibility, we are going to have to define in some
detail basic notions such as gender, number, case, and tense.
32
© John Charles Smith 2015 Language Change and Historical Linguistics
hecho), which has a precise meaning, even though the event time (E) may precede the time of
speaking (S)
(1) Obtuvo su bachillerato el año pasado; por lo tanto, habrá cumplido las condiciones para
ser admitido el año próximo.
follow it
(2) Va a obtener su bachillerato el año próximo; por lo tanto, habrá cumplido las condiciones
para ser admitido el año siguiente.
or be simultaneous with it
(3) Ya tiene su bachillerato; por lo tanto, habrá cumplido las condiciones para ser admitido el
año próximo..
This is a more adequate theory; but it still raises problems. For instance, the future perfect is
represented in three different ways, which should imply it can have three different meanings. But
speakers of Spanish (and English) do not distinguish amongst these meanings (and Comrie (1981)
goes so far as to claim that there is no language in the world that does).
Hornstein, following Comrie, suggests that there is no direct relationship between S and E; this
relationship is always mediated by R. Thus, the first three tenses in the above list would receive the
following descriptions:
Such an account is equivalent to the earlier one in most respects, but has some advantages over it.
For instance, it solves the ‘future perfect problem’ by enabling this tense to be described by the
single formula (E — R) • (S — R), so obviating the need for three separate representations.
33
© John Charles Smith 2015 Language Change and Historical Linguistics
There are also more abstract grounds for preferring this theory, in that it is more constrained —
only two relationships (S and R; E and R) need to be established, as opposed to three (S and R; E
and R; S and E) — and therefore, presumably easier for a child to acquire. (Hardliners would argue
that a three-way system was impossible for a child to acquire...)
Hornstein further suggests that Universal Grammar (broadly speaking, the genetic endowment
which enables a child to acquire a language) automatically associates the reference point (R) and
the event time (E), unless specific information to the contrary is provided. The use of auxiliaries, as
opposed to synthetic verb forms, is seen as one way of ‘flagging’ the marked relationship between
R and E (i.e., that they are not simultaneous). This hypothesis would certainly explain the
distribution of analytic (or ‘compound’) tenses (those involving an auxiliary) and synthetic (or
‘simple’) tenses (those which do not have an auxiliary) in the above table. See also, from a
typological perspective, Dahl (1985:129): ‘[the perfect] is rather consistently marked
periphrastically (about 85 per cent of the cases), the only clear counterexamples being Niger-Congo
languages (Akan, Kikuyu). Typically, constructions involving a copula or some auxiliary together
with some past participle or similar form of the verb are used.’
5.4. The Coseriu typology
Coseriu (1988:213), in a remarkable insight, proposed that the Romance languages were
distinguished from Latin by an iconic typology whereby relational concepts receive relational (i.e.,
analytic) exponence and non-relational concepts receive non-relational (i.e., synthetic) exponence:
‘innere, paradigmatische materielle Bestimmungen für gleichfalls innere, nicht-relationelle
Funktionen und äußere, syntagmatische materielle Bestimmungen für gleichfalls äußere,
relationelle Funktionen.’ The iconicity of this typology is clear: when interpreting the meaning
involves establishing a relationship between two elements at underlying level, the meaning itself is
expressed by two separate items which must be related on the surface; when the meaning does not
require such a relationship to be established, it is expressed by a single item.
Amongst other things, such a typology will account for:
• the general Romance tendency to express number and gender inflectionally but for the
functions of noun case to be replaced by prepositions and word order. To recapitulate our earlier
discussion: Gender is an inherent property of noun types, and hence non-relational. Number is
an inherent property of noun-phrase tokens, and hence also non-relational. The sole function of
case, on the other hand, is to encode the relationship between a noun and some other item in the
sentence (the verb, another noun, a preposition, etc.) — it is therefore a priori relational.
• the use of simple tenses to encode simultaneity of reference point and event (for instance,
the present and the pretérito indefinido) — i.e., non-relationality — and compound tenses to
encode the disjunction of these two points (for instance, the pretérito perfecto and the pretérito
pluscuamperfecto) — i.e., relationality;
• the use of superlative morphology to encode intrinsic size (non-relational), but the use of
a modifier to encode extrinsic size (relational). Thus, una montaña altísima is a mountain which is
just high, high of itself, high full stop; whereas una montaña muy alta is high with respect to some
concrete or abstract point of comparison (another mountain, your average mountain, etc.)
• the use of diminutive morphology to encode intrinsic smallness (non-relational), but the
use of an adjective meaning ‘small’ modifying the noun to encode extrinsic smallness (relational).
Thus, un librito is a book which is small by nature, small in its essence (this is also why the word
can be used hypocoristically, as a term of affection or endearment), whereas un libro pequeño is a
book which is small in relation to or in comparison with some other book or books, or some
yardstick of what one expects a book to look like, etc. (In English, this distinction corresponds
more or less to the lexical difference between ‘little book’ and ‘small book’.)
34
© John Charles Smith 2015 Language Change and Historical Linguistics
6. Two case-studies in grammaticalization:
the pretérito perfecto compuesto and the articles in Spanish
6.1. Grammaticalization
Grammaticalization (see note 1 above) is basically defined as a shift in an item’s value from lexical
to functional, or from less functional to more functional. It therefore by definition involves a
modification of grammatical structure and/or grammatical status: an item is relabelled or
rebracketed, changing its syntactic category (say, from verb to auxiliary) or moving from a
syntactic unit (word) to a morphological unit (affix). In addition, grammaticalization tends to
involve a measure of semantic ‘bleaching’ and may also be accompanied by phonological
reduction. We have already seen instances of such change on p. 5 above, where we mentioned
Vuestra Merced > Usted in Spanish and some other relevant examples
Grammaticalization is a subject of huge debate in contemporary historical linguistics. The standard
works are Traugott & Heine (1991), Hopper & Traugott (2003), and Lehmann (2002). For sober
reflection, see the indispensable papers by Campbell (2001) and Newmeyer (2001). Parapharsig
Campbell, grammaticalization began as a process, became a theory, and is now (at least for some
people) a religion!
A simple but clear (and progressive) example of grammaticalization concerns adverb-formation in
Romance. The ablative case MENTE of the Latin noun MENS ‘mind’ was used literally with an
agreeing (feminine) adjective to mean ‘with an X mind’ — thus CAVTA MENTE ‘with a cautious
mind’. In time, through a process of semantic weakening, this construction came to mean simply
‘in an X way’ — thus CAVTA MENTE ‘in a cautious way’. Subsequently, the original noun became
an ending. But contrast Spanish, where it has become an enclitic, attaching to a phrase, with
French, where it has moved on a stage further and become a derivational (i.e., word-forming)
suffix. Thus, in Spanish, a sequence of adjectives forming a phrase takes only one occurrence of
-mente (lenta y cuidadosamente), whereas in French every adjective requires the suffix -ment
(lentement et précautionneusement). A lexical Latin word has turned into a functional Spanish
clitic and into an even more functional French suffix.
6.2. The pretérito perfecto compuesto
The development of a compound past tense from the HABERE + past participle construction can be
viewed, inter alia, as an example of grammaticalization — the process whereby a (more) lexical
item becomes a (more) functional item.
The ‘standard’ examples of HABERE + past participle discussed in the literature involve invented
sentences such as HABEO EPISTVLAS (or LITTERAS) SCRIPTAS, which allegedly shifted its meaning
from ‘I possess (the) letters which are written’ to ‘I have written (the) letters’. Whilst this statement
of the change of meaning is accurate as far as it goes, the example is a poor one, as it not generally
possible to possess letters which are not written!
Compare:
I’ve got some toasted teacakes — would you like one?
AGENT 1 AGENT 2 toaster unspecified; may or may not be Agent 1
default interpretation: identity of toaster is irrelevant
35
© John Charles Smith 2015 Language Change and Historical Linguistics
A more clear-cut case:
I’ve got a written reply
AGENT 1 AGENT 2 writer is not Agent 1
At the same time, a separate structure, in which the verb ESSE(RE) ‘to be’ was combined with the
past participle, began to be used as the present perfect of intransitive verbs. This had its origins in
the passive perfectum (e.g., AMATVS SVM ‘I have been loved’) — the class of deponent verbs,
which had passive morphology but active meaning (e.g., CONOR ‘I try’, CONATVS SVM ‘I have
tried’), and, more significantly, so-called semi-deponent verbs, which had active morphology in
the infectum but passive morphology in the perfectum (e.g., GAVDEO ‘I rejoice’, GAVISVS SVM ‘I
have rejoiced’) may have been crucial in providing an analogy. For an outstanding survey of all
these developments, see Vincent (1982).
We know that the HABEO + past participle construction must be interpreted as a tense-form when
possessive or causative interpretations are not available — most strikingly, when the verb is
intransitive. Intransitive verbs used ‘absolutely’ (e.g., ‘I have spoken’) may exhibit this
construction by analogy with their transitive use (‘I have spoken many words’, etc.). But once
verbs which can only ever be intransitive (such as ‘elapse’) begin to manifest a HABERE + past
participle form, these occurrences must represent a tense. Likewise, once lexical items which are
incompatible with the notion of possession, such as ‘lose’, ‘forget’, ‘destroy’ are found in this
construction, we know that it cannot be analysed as a possessive.
As Harris & Campbell (1995:30-32) point out, the notion of ‘reanalysis’ is not a new one; it can be
traced back at least as far as the nineteenth century. Compare also the views of two scholars to
whom they acknowledge their debt (ibid.:389n2) — Langacker (1977:58):
[Reanalysis is] change in the structure of an expression or class of expressions that does not involve any
immediate or intrinsic modification of its surface manifestation. Reanalysis may lead to changes at the
surface level [...], but these surface changes can be viewed as the natural and expected result of
functionally prior modifications in rules and underlying representations.
36
© John Charles Smith 2015 Language Change and Historical Linguistics
At first sight, it might seem that Harris & Campbell’s ‘extension’ could be equated with
Timberlake’s ‘actualization’; but Harris & Campbell (1995:80-81) themselves suggest that this is
not in fact the case:
Examining a number of instances of actualization of reanalysis, we have found that each example of
change under actualization [is] itself either an extension or an additional reanalysis. [...] In other
examples of reanalysis we have examined, actualization includes both extension and reanalysis in
morphology, syntax, or lexical items. Just as actualization may involve changes other than extension,
extension may occur other than as a response to reanalysis [for example, it may be induced by language
contact]. Thus, while actualization often consists of extensions, the two are not coextensive.
As a possible example of actualization-as-reanalysis, we may take the following data from the
Romance HABERE + past participle construction under discussion (which is itself mentioned by
Harris & Campbell (1995:182-185) as an example of ‘clause fusion’). As part of this process,
HABERE acquires a second possible category label (Aux, I, T, or whatever, alongside V). One of
the consequences of this development in some Romance languages (Spanish, some southern Italian
dialects) is that the reflex of HABERE becomes restricted to use as an auxiliary (in other words, it
loses the category label V), and is replaced as lexical verb of possession by a reflex of TENERE
(originally ‘to hold’). In Timberlake’s terms, this would be an actualization. In Harris &
Campbell’s terms, the complete ‘delexification’ of <HABERE (i.e., the loss of the category label V)
would arguably count as a further reanalysis (recall that reanalysis is a mechanism which ‘directly
changes underlying structure’, which includes ‘information regarding [...] category labels’ — Harris
& Campbell 1995:61), but one which is contingent on the first reanalysis — hence an actualization,
but not an extension. The change in meaning of <TENERE would then be viewed as an extension of
this further reanalysis.
Implicit in the Harris & Campbell model, then, is a distinction between two types of reanalysis —
‘initial’ reanalysis and ‘actualizing’ reanalysis (my terms).
The reanalysis (the acquisition of a new category label by HABERE) is perceptible semantically (see
the discussion at the end of §1 above), but (by definition) not in surface syntax. Thereafter, there
are (at least) five actualizations/extensions of the reanalysis of HABERE + past participle as a
compound past tense form in Romance (Smith 1989):
6.5. A case-study of actualization (and functionalism?) — the loss of past participle agreement
The disappearance of past participle agreement is differential. Smith (1991; 1993b; 1995b), notes
that the data concerning object-participle agreement in Romance enable us to establish a number of
implicational hierarchies (in the sense of Greenberg 1963), of the form: “if, in a given language or
dialect, the past participle agrees with a direct object of type X, then it will also agree with a direct
object of type Y”. These hierarchies are presented below (the notation X>Y is to be interpreted to
mean that agreement with X implies agreement with Y, but not necessarily vice versa).
37
© John Charles Smith 2015 Language Change and Historical Linguistics
a) Position of Direct Object:
Following > Preceding
b) Identity of Preceding Direct Object:
⎧ Topics ⎫
⎨ Interrogatives ⎬ > Relatives > Clitic Pronouns
⎩ Exclamatives ⎭
c) Person of Clitic Pronoun:
⎧ First Person ⎫
⎨ Second Person ⎬ > Third Person Non-Reflexive
⎩ Third Person Reflexive ⎭
d) Number and Gender of Third-Person Non-Reflexive Clitic Pronoun:
Masculine Plural > Feminine Plural > Feminine Singular
These hierarchies may be read synchronically (as defining patterns of agreement or types of
variation in a given language or dialect) or diachronically (as defining the progression of a change
— i.e., the loss of this type of agreement). On the next page, the resulting patterns of agreement are
illustrated from Catalan, a Romance language which has dialects exemplifying all the above
hierarchies (see Smith 1995a). Sentences exhibiting agreement between past participle and direct
object are in bold type. Significantly, we appear to find no exceptions to these hierarchies and the
resulting patterns of agreement in any Romance language. For further discussion of these data and
their ramifications, including the consequences for a theory of language change, see Smith (1996,
1997, 1999, 2000, 2001, 2002). For a successful application of this hypothesis to Old Spanish, see
Rodríguez Molina (2010).
It can be argued that all of these patterns are ultimately the result of perceptual strategies, and that
the agreement is most resilient when it assists the ‘recovery’ of the direct object in some way. In
this sense, the change is sensitive to ‘functional’ factors (although, as we have already seen, caution
is called for when imputing a change to ‘functionality’). Note that the agreement is in no way
necessary to the recovery of the direct object (comparable sentences involving simple tenses appear
to present no significant parsing problems). However, we are not, of course, arguing that it was
introduced in order to aid sentence-processing, but merely that it survived for longer in those
contexts where it might have such a function. Observe, too, that extension may, as in this case, lag
spectacularly behind reanalysis — in the present instance, by well over a thousand years!
38
© John Charles Smith 2015 Language Change and Historical Linguistics
A. General Agreement (frequent in Old French)
B. Agreement with preceding direct objects (the modern French règle de position)
39
© John Charles Smith 2015 Language Change and Historical Linguistics
6.7. The articles
The first-century A.D. Latin grammarian Quintilian famously states: noster sermo articulos non
desiderat ‘our language has no need of articles’ (Institutio Oratoria 1, 4, 19). However, this did not
mean that it could not express at least something akin to definiteness. A key component of
definiteness is identifiability, which hinges on the crucial notions of ‘knownness’ or
‘presupposedness’. Given that the Latin case-system enabled subject and object to be distinguished
inflectionally in most instances, Latin word order could encode pragmatic functions, such as topic
and focus, much more freely than word order in say, English, where there is no alternative to word
order as an indicator of grammatical function. Although far from identical to definiteness, these
pragmatic notions also involved the crucial notion of ‘knownness’ and/or ‘presupposedness’ and/or
salience.
A suggestive comparison can be drawn with Slavonic languages. Almost all of these have a rich
case-system and flexible word order, rather like Latin. Likewise, they do not have a definite article.
Although we should be careful about using translation as evidence for meaning, it may be
significant that a normal translation of, e.g., Russian Книга на столе (lit. ‘Book [is] on table’) is
‘The book is on the table, whilst На столе книга (lit. ‘On table [is] book’) comes out as ‘A book is
on the table’). The exception to this generalization about Slavonic is Bulgarian (together with its
sister-language, Macedonian), which has lost its case-system, and expresses functions such as
‘subject’ and ‘object’ through word order. It is also the only Slavonic language to have evolved a
definite article. The Bulgarian for ‘a book’ is книга (the bare noun, as in Russian), but ‘the book’ is
книгата, with a suffixal definite article -та.
Another factor to consider is that the emergence of the article, which is a functional category,
parallels the development or extension of other functional categories (prepositions, auxiliaries,
finite complementizers) in Late Latin — it can be seen as part of a more general typological shift
(which we have discussed in earlier lectures).
Finally, it has been suggested that language contact may have played some role in this change.
Although Classical Latin did not have articles, Greek did. (This is doubtless how Quintilian came
to be familiar with the concept of an article; his comment quoted above is an implicit comparison of
Latin and Greek.) When the Greek New Testament was translated into Latin, some of the Greek
definite articles were rendered by Latin demonstratives in what appears to have been a rather
slavish adherence to the original text. Although the Bible will have been familiar to many people
who could not read or write, it is unclear to what extent such literal translations changed everyday
language.
As usual, in as much as the above explanations are not in conflict, they may all be ‘right’ — i.e., the
emergence of the articles of Latin may be due to all of them. We don’t have to choose.
The definite article in Romance evolved from a distal demonstrative (a word for ‘that’). This is an
extremely common source of definite articles (compare Germanic, Malay, etc. — in English, for
instance, the originates as a weakened form of that). Classical Latin had a generic demonstrative IS
and three person-related demonstratives: HIC, ISTE, ILLE. HIC disappeared (possibly as the result of
phonological erosion and perhaps also because it was inflected medially — HIC, HÆC, HOC; HVNC,
HANC, HOC, etc.). ISTE replaced it, and was itself replaced by IPSE (‘the very’, ‘self’). The remote
demonstrative ILLE, which was commonly used cataphorically in Late Latin, is the source of the
definite article in almost all the Romance languages. The ‘second-person’ form IPSE, commonly
used anaphorically, provides the definite article in Sardinian and Balearic Catalan (compare, too,
Gascon place-names such as Sacase < IPSA CASA ‘the house’).
The replacement of HIC by ISTE is regarded as unproblematic by most commentators, but it is not a
straightforward process — why should the second-person form replace the first-person form? In
fact, it is likely that the opposition between the items was inclusive rather than exclusive, with ISTE
encompassing both persons (i.e., the discourse). If this was the case, then the replacement of HIC by
ISTE can be interpreted fairly plausibly as a semantic restriction of the latter.
40
© John Charles Smith 2015 Language Change and Historical Linguistics
A reinforced form of ILLE, preceded by the ostensive particle ECCE (‘lo’, ‘behold’, ‘voilà’) or a
variant of it (*ACCV in Spanish, *ECCV in Italian, ECCE in French, *ACCE in Romanian) was also
available, and this became the unambiguous exponent of remote deixis (‘that’) — compare Spanish
este, ese, aquel).
Subsequently, the ‘definite article’ extended its meaning to encompass generic reference. This
process can be seen as the emergence of what is in part a number and gender marker from an
article, which is a widespread change in the world’s languages (see Greenberg 1978). In addition, it
has been argued independently that generic reference may in some ways be related to definite
reference, as the exhaustive membership of a set is identifiable (see, for instance, Cherchia 1998
and Krifka et al. 1995). Definiteness, as normally understood, entails specificity (the reverse is not
true, as we shall see when we discuss the indefinite article below). However, in some sense,
generic reference might be characterized as definiteness without specificity.
The result of this change is that a contemporary Spanish sentence such as Me gustan las fresas,
although it can bear the meaning ‘I like the strawberries’, has the default interpretation ‘I like
strawberries (in general)’. The former sense can be unambiguously expressed by Me gustan estas
fresas. Observe the parallels with the Latin developments discussed earlier: the ‘bare’
demonstrative comes to have a broadly definite value.
The indefinite article derives from the Latin numeral VNVS ‘one’ (in much the same way as one >
an > a in English, although without the phonological weakening). In Old Spanish it generally refers
only to items with specific (but indefinite) reference (see Elvira 1994, Garachana 2009, Pozas Loyo
2010). (Compare Mary wants to marry a Norwegian — his name is Hans (specific reading of a
Norwegian) and Mary wants to marry a Norwegian — if she can find one (non-specific reading of a
Norwegian). In (modernized) Old Spanish, the first of these sentences would be María quiere
casarse con un noruego... and the second María quiere casarse con noruego... However, like the
definite article, its range is extended, and it comes to mark indefiniteness in general, regardless of
specificity.
This raises the interesting possibility of a parallel between the extension of the definite article and
that of the indefinite article in the history of Spanish (and other Romance languages). Recall our
suggestion above that generic reference might be defined as definiteness without specificity. If this
is the case, then a generalization can be made: both the definite article and the indefinite article are
used in Old Spanish to encode specific reference, but subsequently expand to encompass non-
specific reference as well. Schematically:
specificity specificity
+ – + –
definiteness + le ∅ definiteness + le → le
– un ∅ – un → un
41
© J. C. Smith 2015 Language Change and Historical Linguistics
de a em por
o do ao no pelo
a da à na pela
os dos aos nos pelos
as das às nas pelas
um dum num
uma duma numa
isto disto nisto
este deste neste
esta desta nesta
estes destes nestes
estas destas nestas
isso disso nisso
esse desse nesse
essa dessa nessa
esses desses nesses
essas dessas nessas
aquilo daquilo àquilo naquilo
aquele daquele àquele naquele
aquela daquela àquela naquela
aqueles daqueles àqueles naqueles
aquelas daquelas àquelas naquelas
42
© J. C. Smith 2015 Language Change and Historical Linguistics
43
© J. C. Smith 2015 Language Change and Historical Linguistics
Aitchison, Jean. 2000. Language Change: progress or decay? (third edition). Cambridge: Cambridge
University Press.
Andersen, Henning. 1973. ‘Abductive and deductive change.’ Language 49, 765-793.
Andersen, Henning. 1988. ‘Center and periphery: adoption, diffusion and spread.’ Historical
Dialectology: regional and social, edited by Jacek Fisiak, 39-85. Berlin: Mouton de Gruyter.
Arends, Jacques, Pieter Muysken & Norval Smith.. 1995. Pidgins and Creoles: an introduction.
Amsterdam & Philadelphia, PA: Benjamins.
Bailey, Charles-James & Karl Maroldt. 1977. ‘The French lineage of English.’ Langues en contact –
pidgins – creoles – Languages in Contact, edited by Jürgen Meisel, 21-53. Tübingen: Narr.
Banniard, Michel. 2013. ‘The transition from Latin to the Romance Languages.’ The Cambridge
History of the Romance Languages: Volume II, Contexts, edited by Martin Maiden, John Charles
Smith & Adam Ledgeway, 57-106. Cambridge: Cambridge University Press.
Barber, Charles L. 1997. Early Modern English (new edition). Edinburgh: Edinburgh University Press
(first edition — London: André Deutsch, 1976).
Barnes, Matthew. 1928. ‘Words from the French (-é, -ée).’ Society for Pure English Tract 30, 298-305.
Bartoli, Matteo. 1925. Introduzione alla neolinguistica: principi, scopi, metodi. Genève: Olschki.
Bartoli, Matteo. 1929. ‘La norma linguistica dell’area maggiore.’ Rivista di filologia e d’istruzione
classica 57, 333-345.
Bartoli, Matteo. 1933. ‘Le norme neolinguistiche e la loro utilità per la storia dei linguaggi e dei
costumi.’ Atti della Società italiana per il progresso delle scienze (1933), 157-167.
Bartoli, Matteo. 1945. Saggi di linguistica spaziale. Torino: Bona.
Baugh, Albert C. 1935. ‘The chronology of French loan-words in English.’ Modern Language Notes 50,
90-93.
Bentley, Harold W. 1932. A Dictionary of Spanish Terms in English, with special reference to the
American Southwest. New York, NY: Columbia University Press.
Berndt, Rolf. 1965. ‘The linguistic situation in England from the Norman Conquest to the loss of
Normandy.’ Philologica Pragensia 8, 145-163.
Berndt, Rolf. 1976. ‘French and English in thirteenth-century England: an investigation into the
linguistic situation after the loss of the Duchy of Normandy and other continental dominions.’ Aspekte
der anglistischen Forschung in der DDR: Martin Lehnert zum 65. Geburtstag (Sitzungsberichte der
Akademie der Wissenschaften der DDR: Gesellschaftswissenschaften, 1976/1), 129-150. Berlin:
Akademie-Verlag.
Bickerton, Derek. 1991. Roots of Language. Ann Arbor, MI: Karoma.
Binnick, Robert I. 1991. Time and the Verb: a guide to tense and aspect. Oxford: Oxford University
Press.
Blake, Norman. 1992. ‘Introduction.’ The Cambridge History of the English Language: volume II,
1066-1476, edited by Norman Blake, 1-22. Cambridge: Cambridge University Press.
Bliss, A. J. 1966. A Dictionary of Foreign Words and Phrases in Current English. London: Routledge
& Kegan Paul.
Bonfante, Giulio. 1947. ‘The neolinguistic position’, in Language 23 (1947), pp. 344-375, esp. pp. 368-
374 (also at http://www.jstor.org/stable/410296).
Bradbury, Malcolm. 1983. Rates of Exchange. London: Secker & Warburg.
Brincat, G. 1991. ‘Spigolando nel Sunday Times: parole italiane e pseudoitalianismi nell’inglese di
oggi.’ L’italiano allo specchio: aspetti dell’italianismo recente (Atti del Primo Convegno della Società
Internazionale di Linguistica e Filologia Italiana, Siena, 28-31 marzo 1989, volume 2˚), edited by
Lorenzo Coveri, 7-14. Torino: Rosenberg e Sellier.
Brown, Keith, ed. 2005. Encyclopedia of Language and Linguistics (second edition) (14 vols.). Oxford:
Elsevier.
Burnley, David. 1992. ‘Lexis and semantics.’ The Cambridge History of the English Language: volume
II, 1066-1476, edited by Norman Blake, 409-499. Cambridge: Cambridge University Press.
Bynon, Theodora. 1977. Historical Linguistics. Cambridge: Cambridge University Press.
Caluwé Dor, J. 1983. ‘Chaucer’s contribution to the English vocabulary: a chronological survey of
French loan-words.’ North-West European Language Evolution 2, 73-91.
44
© J. C. Smith 2015 Language Change and Historical Linguistics
Campbell, Lyle. 2001. What’s wrong with grammaticalization? Language Sciences 23, 113-161.
Campbell, Lyle. 2013. Historical Linguistics: an introduction (third edition). Edinburgh: Edinburgh
University Press. (US edition: Cambridge, MA: MIT Press, 1999.)
Campbell, Lyle & Mauricio Mixco. 2007. Glossary of Historical Linguistics. Edinburgh: Edinburgh
University Press.
Cannon, Garland. 1996. ‘Recent borrowings from Spanish into English.’ Spanish Loanwords in the
English Language: a tendency towards hegemony reversal, edited by Félix Rodríguez González, 41-
60. The Hague: Mouton de Gruyter.
Caroll, John B., Peter Davies & Barry Richman. 1971. The American Heritage Word Frequency Book.
New York, NY: American Heritage.
Chambers, J. K. & Peter Trudgill. 1998. Dialectology (second edition). Cambridge: Cambridge
University Press.
Cheshire, Jenny, Paul Kerswill, Susan Fox & Eivind Torgersen. 2011. ‘Contact, the feature pool and the
speech community: the emergence of Multicultural London English.’ Journal of Sociolinguistics 15,
151–196.
Chierchia, Gennaro. 1998. ‘Reference to kinds across languages.’ Natural Language Semantics, 6, 339-
405
Chirol, Laure. 1973. Les « Mots français » et le mythe de la France en anglais contemporain. Paris:
Klincksieck.
Clackson, James & Geoffrey Horrocks. 2007. The Blackwell History of the Latin Language. Oxford:
Blackwell.
Coetsem, Frans van. 1988. Loan Phonology and the Two Transfer Types in Language Contact.
Dordrecht: Foris.
Coetsem, Frans van. 2000. A General and Unified Theory of the Transmission Process in Language
Contact. Heidelberg: Winter.
Coleman, J. M. 1995. ‘The chronology of French and Latin loan words in English.’ Transactions of the
Philological Society 93, 95-124.
Comrie, Bernard. 1976. Aspect. Cambridge: Cambridge University Press.
Comrie, Bernard. 1981. ‘On Reichenbach’s approach to tense.’ Papers from the Seventeenth Regional
meeting of the Chicago Linguistic Society, edited by Roberta A. Hendrick, Carrie S. Masek & Mary
Frances Miller, 24-30. Chicago: University of Chicago.
Comrie, Bernard. 1985. Tense. Cambridge: Cambridge University Press.
Coseriu, Eugenio. 1988. ‘Der romanische Sprachtypus: Versuch einer neuen Typologisierung der
romanischen Sprachen.’ Energeia und Ergon: Sprachliche Variation, Sprachgeschchte,
Sprachtypologie. Band I, Schriften von Eugenio Coseriu. (1965-1987), edited by Jörn Albrecht, 207-
224. Tübingen: Narr.
Croft, William. 2000. Explaining Language Change: an evolutionary approach. London & New York,
NY: Longman.
Crowley, Terry & Claire Bowern. 2010. An Introduction to Historical Linguistics (fourth edition).
Oxford: Oxford University Press.
Crystal, David. 2008. A Dictionary of Linguistics and Phonetics (sixth edition). Chichester & Malden,
MA.: Wiley–Blackwell.
Crystal, David. 2010. The Cambridge Encyclopedia of Language (revised edition). Cambridge:
Cambridge University Press.
Culpeper, Jonathan & Phoebe Clapham. 1996. ‘The borrowing of Classical and Romance words into
English: a study based on the electronic Oxford English Dictionary.’ International Journal of Corpus
Linguistics 1, 199-216.
Dahl, Östen. 1984. ‘Temporal distance: remoteness distinctions in tense-aspect systems.’ Explanations
for Language Universals, edited by Brian Butterworth, Bernard Comrie & Östen Dahl, 105-122.
Berlin, New York, NY & Amsterdam: Mouton.
Dahl, Östen. 1985. Tense and Aspect Systems. Oxford: Blackwell.
Dalton-Puffer, Christiane. 1995. ‘Middle English is a creole and its opposite: on the value of plausible
speculation.’ Linguistic Change under Contact Conditions, edited by Jacek Fisiak, 35-52. Berlin &
New York, NY: Mouton de Gruyter.
45
© J. C. Smith 2015 Language Change and Historical Linguistics
Dalton-Puffer, Christiane. 1996. The French Influence on English Morphology: a corpus-based study of
derivation. Berlin & New York, NY: Mouton de Gruyter.
Dawkins, Richard. 1996. Climbing Mount Improbable. London: Viking. (Paperback edition,
Harmondsworth: Penguin, 1997).
Dawson, Hope C. 2003. ‘Defining the Outcome of Language Contact: Old English and Old
Norse.’ Ohio State University Working Papers in Linguistics 57, 40-57.
Dekeyser, Xavier. 1986. ‘Romance loans in Middle English: a reassessment.’ Linguistics across
Historical and Cultural Boundaries — In Honour of Jacek Fisiak on the Occasion of his Fiftieth
Birthday: volume I, Linguistic Theory and Historical Linguistics, edited by Dieter Kastovsky &
Aleksander Szwedek, 253-265. Berlin: Mouton de Gruyter.
Dekeyser, Xavier & Luc Pauwels. 1990. ‘The demise of the Old English heritage and lexical innovation
in Middle English: two intertwined developments.’ Leuvense Bijdragen 79, 1-23.
Deutscher, Guy. 2002. ‘On the misuse of the notion of “abduction” in linguistics.’ Journal of
Linguistics 38, 469-485.
Domingue, Nicole Z. 1977. ‘Middle English: another creole?’ Journal of Creole Studies 1, 89-100.
Donohue, Mark & John Charles Smith. 1998. ‘What’s happened to us? Some developments in the
Malay pronoun system.’ Oceanic Linguistics 37, 65-84.
Durkin, Philip. 2014. Borrowed Words: a history of loanwords in English. Oxford: Oxford University
Press.
Eckert, Penelope. 1989. Jocks and Burnouts: social categories and identity in the High School. New
York, NY: Teachers’ College Press.
Eckert, Penelope. 2000. Linguistic Variation as Social Practice: the linguistic construction of identity in
Belten High. Oxford: Blackwell.
Elvira, Javier. 1994. ‘Uno en el español antiguo.’ Verba 21, 167-182.
Ferguson, Charles. 1959. ‘Diglossia.’ Word 15, 325-240.
Fischer, Olga. 1989. ‘The origin and spread of the accusative and infinitive construction in English.’
Folia linguistica historica 8, 143-217.
Fishman, Joshua A. 1967. ‘Bilingualism with and without diglossia, diglossia with and without
bilingualism.’ Journal of Social Issues 23, 29-38.
Fleischman, Suzanne. 2000. ‘Methodologies and ideologies in historical linguistics: on working with
older languages.’ Textual Parameters in Older Languages, edited by Susan C. Herring, Pieter van
Reenen & Lene Schøsler, 33-58. Amsterdam & Philadelphia, PA: John Benjamins.
Foulet, Alfred & Mary Blakely Speer. 1979. On Editing Old French Texts. Lawrence, KS: The
Regents’ Press of Kansas.
Francis, W. Nelson & Henry Kučera. 1982. Frequency Analysis of English Usage: lexicon and
grammar. Boston, MA: Houghton Mifflin.
Frawley, William J. 2003. International Encyclopedia of Linguistics (second edition). Oxford: Oxford
University Press.
Garachana, Mar. 2009. ‘Creación y generalización del artículo indefinido.’ Sintaxis histórica de la
lengua española: la frase nominal, edited by Concepción Company Company, I, 387-464. Mexico DF:
FCE; UNAM).
García González, Javier. 1993. ‘El contacto de dos lenguas: los arabismos en el español medieval y en la
obra alfonsí.’ Cahiers de linguistique hispanique médiévale 18, 335-365.
Gerritsen, Marinel. 1999. ‘Divergence of dialects in a linguistic laboratory near the Belgian–Dutch–
German border: similar dialects under the influence of different standard languages.’ Language
Variation and Change 11, 43-65, available online at:
http://marinelgerritsen.files.wordpress.com/2011/12/gerritsen-1999.pdf.
Gooch, Anthony. 1996. ‘Aspects of the incidence of Hispanicisms in British English: a study in
language, history, politics and psychology.’ Spanish Loanwords in the English Language: a tendency
towards hegemony reversal, edited by Félix Rodríguez González, 231-251. The Hague: Mouton de
Gruyter.
Görlach, Manfred. 1991. Introduction to Early Modern English. Cambridge: Cambridge University
Press (updated English version of Einführung ins Frühneuenglische. Heidelberg: Quelle & Meyer,
1978).
46
© J. C. Smith 2015 Language Change and Historical Linguistics
Greenberg, Joseph H. 1963. ‘Some universals of grammar with particular reference to the order of
meaningful elements.’ Universals of Language (report of a conference held at Dobbs Ferry, New
York, April 13-15, 1961), edited by Joseph H. Greenberg, 58-90. Cambridge, MA: MIT Press.
Greenberg, Joseph H. 1978. ‘How does a language acquire gender markers?’, in Joseph H. Greenberg
(ed.) Universals of Human Language: volume 3, Word Structure, edited by Joseph H. Greenberg, 47-
82. Stanford, CA: Stanford University Press.
Hall, Kathleen Currie. 2013. ‘A typology of intermediate phonological relationships.’ The Linguistic
Review 30, 215-275.
Hall, Robert A. jr. 1946. ‘Bartoli’s “neolinguistica”.’ Language 22, 273-283 and at
http://www.jstor.org/stable/409919.
Harris, Alice C. & Lyle Campbell. 1995. Historical Syntax in Cross-Linguistic Perspective. Cambridge:
Cambridge University Press.
Harris, Martin. 1980a. ‘The marking of definiteness in Romance.’ Historical Morphology, edited by
Jacek Fisiak, 301-314. Berlin: Mouton de Gruyter.
Harris, Martin. 1980b. ‘The marking of definiteness: a diachronic perspective.’ Papers from the 4th
International Conference on Historical Linguistics, edited by Elizabeth C. Traugott, Rebecca LaBrun
& Susan Shepherd, 75-86. Amsterdam & Philadelphia, PA: John Benjamins.
Hock, Hans Henrich [sic]. 1991. Principles of Historical Linguistics (second edition, revised and
updated). Berlin & New York, NY: Mouton de Gruyter.
Hofland, Knut & Stig Johansson. 1982. Word Frequencies in British and American English. Bergen:
Norwegian Computing Centre for the Humanities.
Hopper, Paul J. & Elizabeth Closs Traugott. 2003. Grammaticalization (second edition). Cambridge:
Cambridge University Press.
Hornstein, Norbert. 1990. As Time Goes By: tense and Universal Grammar. Cambridge, MA: MIT
Press.
Janson, Tore. 1991. ‘Language change and metalinguistic change: Latin to Romance and other cases.’
Latin and the Romance Languages in the Early Middle Ages, edited by Roger Wright, 19-28. London
& New York, NY: Routledge (second edition, University Park, PA: Pennsylvania State University
Press, 1996); in addition, most of this text is available online at http://books.google.co.uk/books?id=-
bIOAAAAQAAJ&pg=PA19.
Jeffers, Robert J. & Lehiste, Ilse. 1979. Principles and Methods for Historical Linguistics. Cambridge,
MA: MIT Press.
Jochnowitz, George. 1973. Dialect Boundaries and the Question of Franco-Provençal. Den Haag:
Mouton.
Käsmann, Hans. 1958. ‘Zur Rezeption französischer Lehnwörter im Mittelenglischen.’ Anglia 76, 285-
298.
Kerswill, Paul. 2015. ‘Sociolinguistic typology, dialect formation and dialect levelling in industrial and
post-industrial Britain: vernacular speech since 1800.’ MS, University of York.
Kibbee, Douglas A. 1991. For to Speke French Trewely: the French language in England, 1000-1600;
its status, description and instruction. Amsterdam & Philadelphia, PA: John Benjamins.
Kiparsky, Paul. 2008. ‘Universals constrain change; change results in typological generalizations.’
Linguistic Universals and Language Change, edited by Jeff Good, 23-53. Oxford: Oxford University
Press.
Krifka, Manfred, et al. 1995. ‘Genericity: an introduction.’ The Generic Book, edited by Gregory N.
Carlson & Francis Jeffry Pelletier, 1-24. Chicago & London: University of Chicago Press.
Kroch, Anthony. 1989. ‘Reflexes of grammar in patterns of language change’, Language Variation and
Change 1, 199-244 (downloadable from:
ftp://babel.ling.upenn.edu/papers/faculty/tony_kroch/papers/reflexes.pdf).
Kroch, Anthony. 2001. ‘Syntactic Change’. The Handbook of Contemporary Syntactic Theory, edited
by Mark Baltin & Chris Collins, 699-729. Oxford: Blackwell (downloadable from:
ftp://babel.ling.upenn.edu/papers/faculty/tony_kroch/papers/diachronic-syntax-99.pdf).
Labov, William. 1963. ‘The social motivation of a sound change.’ Word 19, 273-309. (Reprinted in
William Labov Socio-linguistic Patterns. Philadelphia, PA: University of Pennsylvania Press, 1972,
1-42.)
47
© J. C. Smith 2015 Language Change and Historical Linguistics
Labov, William. 1994. Principles of Linguistic Change. Volume 1: Internal Factors. Oxford:
Blackwell.
Labov, William. 2001. Principles of Linguistic Change. Volume 2: Social Factors. Oxford: Blackwell.
Labov, William. 2007. ‘Transmission and diffusion.’ Language 83, 344-387; downloadable from
http://www.ling.upenn.edu/~wlabov/Papers/TD.pdf
Labov, William. 2010. Principles of Linguistic Change. Volume 3: Cognitive and Cultural Factors.
Chichester & Malden, MA: Wiley–Blackwell.
Ladefoged, Peter. 1997. ‘Sounds of the world’s languages.’ MS, University of California Los Angeles.
Langacker, Ronald W. 1977. ‘Syntactic reanalysis.’ Mechanisms of Syntactic Change, edited by Charles
N. Li, 57-139. Austin & London: University of Texas Press.
Lass, Roger. 1990. ‘How to do things with junk: exaptation in language evolution.’ Journal of
Linguistics 26, 79-102.
Lass, Roger. 1992. ‘Phonology and morphology.’ The Cambridge History of the English Language:
volume II, 1066-1476, edited by Norman Blake, 23-155. Cambridge: Cambridge University Press.
Lass, Roger. 1997. Historical Linguistics and Language Change. Cambridge: Cambridge University
Press.
Lass, Roger. 2000. ‘Periodization in English and the concept “Middle”.’ Placing Middle English in
Context, edited by Irma Taavitsainen, Terttu Nevalainen, Päivi Pahta & Matti Rissanen, 7-42. Berlin
& New York, NY: Mouton de Gruyter.
Ledgeway, Adam. 2012. From Latin to Romance: morphosyntactic typology and change. Oxford:
Oxford University Press.
Lehmann, Christian. 2002. Thoughts on Grammaticalization (second, revised edition). Arbeitspapiere
des Seminars für Sprachwissenschaft der Universität Erfurt 9, available on line at
http://www.christianlehmann.eu/publ/ASSidUE09.pdf
Lehmann, Winfred P. 1992. Historical Linguistics: an introduction (third edition). London: Routledge.
Lewontin, R. C. 1990. ‘How much did the Brain have to Change for Speech?’ Behavioral and Brain
Sciences 13, 740-741.
Lieberson, Stanley. 2000. A Matter of Taste: how names, fashions and culture change. New Haven, CT:
Yale University Press.
Lightfoot, David. 1991. How to Set Parameters: arguments from language change. Cambridge, MA:
MIT Press.
Lightfoot, David. 1999. The Development of Language: acquisition, change, and evolution. Oxford:
Blackwell.
Lloyd, Paul M. 1987. From Latin to Spanish: historical phonology and morphology of the Spanish
Language. Phildelphia, PA: American Philosophical Society.
Lloyd, Paul. M. 1991. ‘On the names of languages (and other things).’ Latin and the Romance
Languages in the Early Middle Ages, edited by Roger Wright, 9-18. London & New York, NY:
Routledge (second edition, University Park, PA: Pennsylvania State University Press, 1996); in
addition, most of this text is available online at
http://books.google.co.uk/books?id=-bIOAAAAQAAJ&pg=PA19.
Lodares, Juan Ramón. 1996. ‘Hispanic tracks in English: cowboys and gold-rushers in the Old West.’
Spanish Loanwords in the English Language: a tendency towards hegemony reversal, edited by Félix
Rodríguez González, 157-176. The Hague: Mouton de Gruyter.
Lyons, Christopher. 2000. ‘The origins of definiteness marking’, (eds.) Historical Linguistics 1995:
selected papers from the 12th International Conference on Historical Linguistics, Manchester, August
1995. Volume 1: General issues and non-Germanic languages, edited by John Charles Smith & Delia
Bentley, 223-241. Amsterdam & Philadelphia, PA: John Benjamins.
Mackenzie, Fraser. 1939. Les Relations de l’Angleterre et de la France d’après le vocabulaire. Paris:
Droz.
Mańczak, Witold. 1965. ‘La nature des archaïsmes des aires latérales.’ Lingua 13, 177-184.
Mańczak, Witold. 1976. ‘Le cinquantième anniversaire des “normes” de Bartoli.’ General Linguistics
16, 1-8.
Marshall, Jonathan. 2004. Language Change and Sociolinguistics: rethinking social networks.
Basingstoke: Palgrave Macmillan.
Matthews, P. H. 2014. The Concise Oxford Dictionary of Linguistics. Oxford: Oxford University Press.
48
© J. C. Smith 2015 Language Change and Historical Linguistics
49
© J. C. Smith 2015 Language Change and Historical Linguistics
Røyneland, Unn. 2004. Dialektnivellering, ungdom og identitet: ein komparativ analyse av språkleg
variasjon og endring i to tilgrensande dialektområde, Røros og Tynset. Unpublished doctoral thesis,
Universitet i Oslo.
Ruhlen, Merritt. 1994. The Origin of Language. New York, NY: Wiley.
Sandfeld, Kristian. 1930. Linguistique balkanique: problèmes et résultats. Paris: Champion.
Schleicher, August. 1853. ‘Die ersten Spaltungen des indogermanischen Urvolkes.’ Allgemeine
Zeitschrift fuer Wissenschaft und Literatur, 786-787.
Schleicher, August. 1863. Die Darwinsche Theorie und die Sprachwissenschaft: offenes Sendschreiben
an Herrn Dr Ernst Hackel. Weimar: Böhlau.
Schleicher, August. 1865. Die Bedeutung der Sprache für die Naturgeschichte des Menschen. Weimar:
Böhlau.
Schmidt, Johannes. 1872. Die Verwandtschaftsverhältnisse der indogermanischen Sprachen. Weimar:
Böhlau.
Schuchardt, Hugo. 1885. Ueber die Lautgesetze: gegen die Junggrammatiker. Berlin: Oppenheim.
Serjeantson, Mary S. 1935. A History of Foreign Words in English. London: Routledge & Kegan Paul.
Shelly, Percy van Dyke. 1952. English and French in England 1066-1100. Philadelphia, PA: University
of Pennsylvania Press.
Siegel, Jeff. 1985. ‘Koines and koineization.’ Language in Society 14, 357-378 (also online at
http://www.jstor.org/stable/4167665.
Slettengren, Emrik. 1932. Contributions to the Study of French Loanwords in Middle English. Örebro:
Lindhska.
Smith, John Charles. 1989. ‘Actualization reanalyzed: evidence from the Romance compound past
tenses.’ Synchronic and Diachronic Approaches to Linguistic Variation and Change (Georgetown
University Round Table on Languages and Linguistics 1988), edited by Thomas J. Walsh, 310-325.
Washington, DC: Georgetown University Press.
Smith, John Charles. 1991. ‘Problemi dell’accordo del participio passato coll’oggetto diretto nei tempi
composti coniugati con avere in italiano, con speciale riferimento ai dialetti.’ Tra Rinascimento e
strutture attuali: saggi di linguistica italiana (Atti del Primo Convegno della Società Internazionale
di Linguistica e Filologia Italiana, Siena, 28-31 marzo 1989, volume 1˚), edited by Luciano Giannelli,
Nicoletta Maraschio, Teresa Poggi Salani & Massimo Vedovelli, 365-371. Torino: Rosenberg &
Sellier.
Smith, John Charles. 1993a. ‘The agreement of the past participle conjugated with avoir and a preceding
direct object: a brief history of prescriptive attitudes.’ Authority and the French Language: papers
from a conference at the University of Bristol, edited by Rodney Sampson, 87-125. Münster: Nodus.
Smith, John Charles. 1993b. ‘La desaparición de la concordancia entre participio de pasado y objeto
directo en castellano y catalán: aspectos geográficos e históricos.’ Actas del Primer Congreso Anglo-
Hispano: tomo I, Lingüística, edited by Ralph Penny, 275-285. Madrid: Castalia..
Smith, John Charles. 1995a. ‘Agreement between past participle and direct object in Catalan: the
hypothesis of Castilian influence revisited.’ Linguistic Change under Contact Conditions, edited by
Jacek Fisiak, 271-289. Berlin & New York, NY: Mouton de Gruyter (Trends in Linguistics —
Studies and Monographs 81).
Smith, John Charles. 1995b. ‘Perceptual factors and the disappearance of agreement between past
participle and direct object in Romance.’ Linguistic Theory and the Romance Languages, edited by
John Charles Smith & Martin Maiden, 161-180. Amsterdam & Philadelphia, PA: John Benjamins
(Current Issues in Linguistic Theory 122).
Smith, John Charles. 1996. ‘Surfonctionnalité et hyperanalyse: l’accord du participe passé dans les
langues romanes à la lumière de deux théories récentes.’ Faits de langues 8, 113-120.
Smith, John Charles. 1997. ‘Types and tokens in language change: some evidence from Romance.’
Language History and Linguistic Modelling: a Festschrift for Jacek Fisiak on his 60th birthday,
edited by Raymond Hickey & Stanisław Puppel, 1099-1111. Berlin & New York, NY: Mouton de
Gruyter (Trends in Linguistics — Studies and Monographs 101).
Smith, John Charles. 1999. ‘Markedness and morphosyntactic change revisited: the case of Romance
past participle agreement.’ The Emergence of the Modern Language Sciences: studies on the
transition from historical-comparative to structural linguistics in honour of E. F. K. Koerner.
50
© J. C. Smith 2015 Language Change and Historical Linguistics
Volume 2: Methodological Perspectives and Applications, edited by Sheila Embleton, John E. Joseph
& Hans-Josef Niederehe, 203-215. Amsterdam & Philadelphia, PA: John Benjamins.
Smith, John Charles. 2000. ‘Reanalysis and extension: notes and queries.’ Linguistics and Language
Studies: exploring language from different perspectives, edited by Irena Kovačič, Milena Milojević-
Sheppard, Silvana Orel-Kos & Janez Orežnik, 45-60. Ljubljana: Filozofska Fakulteta Univerze v
Ljubljani.
Smith, John Charles. 2001. ‘Markedness, functionality, and perseveration in the actualization of a
morphosyntactic change.’ Actualization: linguistic change in progress, edited by Henning Andersen,
203-223. Amsterdam & Philadelphia, PA: John Benjamins, 2001 (Current Issues in Linguistic
Theory 219).
Smith, John Charles. 2002. ‘L’accordo fra participio passato e oggetto diretto in qualche varietà romanza
e le sue implicazioni per una teoria del cambiamento sintattico’, in Lesser-Used Languages and
Romance Linguistics, edited by Tullio De Mauro & Shigeaki Sugeta, 169-184. Roma: Bulzoni.
Smith, John Charles. 2005. ‘Some refunctionalizations of the nominative-accusative opposition between
Latin and Gallo-Romance.’ A Companion in Linguistics: a Festschrift for Anders Ahlqvist on the
occasion of his sixtieth birthday, edited by Bernadette Smelik, Rijcklof Hofman, Camiel Hamans &
David Cram, 269-285. Nijmegen: De Keltische Draak.
Smith, John Charles. 2006. ‘How to do things without junk: the refunctionalization of a pronominal
subsystem between Latin and Romance.’ New Perspectives on Romance Linguistics: selected papers
from the 35th Linguistic Symposium on Romance Languages (LSRL), Austin, Texas, February 2005.
Volume II: Phonetics, Phonology and Dialectology, edited by Jean-Pierre Y. Montreuil, 183-205.
Amsterdam & Philadelphia, PA: John Benjamins.
Smith, John Charles. 2008. ‘The refunctionalisation of first-person plural inflection in Tiwi.’
Morphology and Language History: in honour of Harold Koch, edited by Claire Bowern, Bethwyn
Evans & Luisa Miceli, 341-348. Amsterdam & Philadelphia, PA: John Benjamins.
Smith, John Charles. 2011. ‘Change and continuity in form–function relationships.’ The Cambridge
History of the Romance Languages. Volume 1: Structures, edited by Martin Maiden, John Charles
Smith & Adam Ledgeway, 268-317. Cambridge: Cambridge University Press.
Steiner, George. 1971. Extraterritorial: papers on literature and the language revolution. New York,
NY: Atheneum.
Studer, Paul & E. G. R. Waters. 1924. Historical French Reader: Medieval Period. Oxford: Clarendon
Press.
Sturtevant, Edgar H. 1947. Introduction to Linguistic Science. New Haven, CT: Yale University Press.
Sweetser, Eve. 1990. From Etymology to Pragmatics: metaphorical and cultural aspects of semantic
structure. Cambridge: Cambridge University Press.
Thomason, Sarah Grey & Terrence Kaufman. 1988. Language Contact, Creolization, and Genetic
Linguistics. Berkeley, CA, etc.: University of California Press.
Thomason, Sarah G. 2001. Language Contact: an introduction. Edinburgh: Edinburgh University Press
Thorndike, Edward L. & Irving Lorge. 1944. The Teacher’s Word Book of 30,000 Words. New York,
NY: Teachers’ College of Columbia University.
Timberlake, Alan. 1977. ‘Reanalysis and actualization in syntactic change.’ Mechanisms of Syntactic
Change, edited by Charles N. Li, 141-177. Austin & London: University of Texas Press.
Toon, Thomas E. 1983. The Politics of Early Old English Sound Change. New York, NY & London:
Academic Press.
Trask, R. L. 2007. Trask’s Historical Linguistics, revised by Robert McCall Millar. London: Edward
Arnold.
Trask, R. L. 2000. The Dictionary of Historical and Comparative Linguistics. Edinburgh: Edinburgh
University Press.
Traugott, Elizabeth Closs. 1982. ‘From propositional to textual and expressive meaning: some
semantic-pragmatic aspects of grammaticalization.’ Perspectives on Historical Linguistics, edited by
Winfred P. Lehmann & Yakov Malkiel, 245-271. Amsterdam & Philadelphia, PA: John Benjamins.
Traugott, Elizabeth Closs. 1989. ‘On the rise of epistemic meanings in English: an example of
subjectification in semantic change.’ Language 65, 31-55.
Traugott, Elizabeth Closs. 2003. ‘From subjectification to intersubjectification.’ Motives for Language
51
© J. C. Smith 2015 Language Change and Historical Linguistics
52