On the Typological Evolution
of the Danish Morphological System
Prof. Vulf Plotkin, D.Sc.
ABSTRACT: The highly analyticized Danish language has developed a
morphological feature found in other formerly synthetic languages undergoing
analyticization, namely composite word-forms, in which the auxiliary and the
lexical element are joined as the two parts of a grammatical compound. Such
compounds are built in Danish with the postfixed definite article -en/-et
and the Present tense marker -er, which explains the retention of the
stød in their first components. Like any participant in a compound, these
auxiliaries are also used as separate words, namely the indefinite article en/et
and the link-verb er, respectively. The positionally determined
variability of the correlated grammatical meanings conveyed by these
auxiliaries is in full accordance with the basic features of analytical
typology.
1. BACKGROUND
The evolution of both Germanic and Romance languages in the past two
millennia was determined by a typological trend towards analyticization. Given
the number and diversity of the languages involved and the vastness of the area
they cover, this trend has naturally found a variety of linguistic
manifestations, some of which characterize each of the two language groups as a
whole, while others are confined to certain groupings within them or to
individual languages.
An example can be seen in the establishment of fixed word-order, which,
being among the primary features of analyticization, is common to the languages
in both groups, but is realized in them in different ways. While word-order in
a predominantly synthetic language can be entirely free from grammatical
restrictions, permitting every rearrangement within the sentence, it can never
be fully fixed in a natural language and must allow at least some syntactic
rearrangements. The measure of permitted word-order freedom and the kind of
grammatical mechanism for its realization are therefore important features in
the typological characterization of a language with basically fixed word-order.
In that respect the Germanic and Romance languages display three solutions. All
the Romance languages have developed a common strategy to provide for a wide
range of sentence rearrangements compatible with fixed word-order,
characterized by the use of special word-order shifters of pronominal nature
affixed to the verb. Incidentally, this strategy has also spread to the
adjacent non-Romance languages in the Balkan area, viz. Greek, Bulgarian and
Albanian, which makes it in fact an areal feature. Most Germanic languages have
achieved a measure of word-order freedom by placing the shifted component into
a position at the beginning of the sentence specially designated for that
purpose. But this solution cannot be described as common Germanic, because a
very different mechanism has emerged in the English language for syntactic
rearrangements permissible under its specific rules of fixed word-order.
There are also significant differences, both between the two language
groups and within them, concerning dynamic interaction between the
diametrically opposite typological trends of analyticization and
syntheticization. Those trends are in fact vectors directed towards the two
poles of a typological continuum, in which both forces are never completely at
rest. In a predominantly synthetic language analytical formations will emerge
from time to time, but they might fail to gain sufficient ground in the
synthetic paradigms and will then undergo gradual erosion, ending up as
inflected forms with the auxiliary reduced to an affix. A strong trend towards
re-syntheticization is sometimes observed even in a language that has gone far
in abandoning its former synthetic features and has reached a rather high
degree of analyticization.
The typological evolution of the three major language groups in Europe -
Germanic, Romance and Slavic - presents essentially different pictures of the
dynamic interaction between the two trends. The Slavic languages have largely
preserved their inherited synthetic typology despite the emergence in all of
them of analytical verb-forms for the Future, the Perfect and the Reflexive, as
well as definite forms of adjectives with pronouns as auxiliaries. Few of these
analytical forms still retain that status; while others have already been
syntheticized, either completely with the former auxiliary becoming an
inflection or being simply dropped, or incompletely if the auxiliary, while
preserving certain features of a separate word, has lost some others.
Analyticization in the Romance group started as early as in Latin, whose
descendants inherited a basically analytical grammar with a greatly diminished
inventory of synthetic forms. But all of them have displayed a strong trend
towards re-syntheticization, as a result of which many originally analytical
formations are now simplified with the former auxiliaries reduced to
inflections, e.g. Lat. cantare habeo > Fr. chanterai, Ital. canteró,
Sp. cantaré ‘will sing’.
Re-syntheticization, however, is by no means a simple one-way process of
auxiliary words gradually turning into inflectional morphemes. Between these
two extremes there are several significant intermediate stages. At the first
one the auxiliary is still a separate word formally, but it is being deprived
of its former status with the loss of a number of lexical and syntactic
properties of a full-fledged word. Examples of this stage abound in the Romance
and Slavic languages of Central and South-Eastern Europe. For instance, the
Perfect of the Romanian verbs employs an auxiliary resulting from the reduction
of Latin habere: am/ai/a/am/aţi/au purtat ‘have carried’.
The formation is still apparently analytical, but the auxiliary (which has
diverged from the separate word am/ai/are/avem/aveţi/au
‘have’ in three of its six forms of person and number) is severely restricted
in its syntactic properties: unlike its counterparts in the Perfect forms of
French, English, German and many other languages, it is practically
inseparable, with very few exceptions, from the adjacent lexical component of
the analytical form.
The next stage in the reduction of an auxiliary is its merger with the
lexical component into one word without, however, becoming an inflection. The
process of lexical simplification, i.e. amalgamation of a word combination like
Old English dæges eage into Modern English daisy, could
not bypass the intermediate compound stage of day's-eye, at which each
of the two components was clearly related to the respective separate word.
Similarly, an analytical combination may pass through the stage of a
grammatical compound before becoming a simplified inflected form. At that stage
the auxiliary element, while already inseparable from its lexical counterpart,
must be correlated with a separate word and function as its representation in
the composite (not yet simply inflected!) grammatical form. That stage can be
illustrated by the Plusquamperfectum of Latin verbs: portav-era-m/s/t/mus/tis/nt
‘have carried’, which is a composite form consisting of the Perfect base portav-
of the lexical element portare ‘carry’ and the auxiliary verb esse
‘be’ in its inflected forms of tense, person and number.
But grammatical compounds are not necessarily products of the
syntheticization of analytical formations. When composition is firmly established
in a language as a way of building paradigm forms, grammatical compounds may
emerge from the opposite direction, namely, from former inflections
reinterpreted within the typologically transformed language system. Beside the
above-mentioned Perfect the Romanian verb has an Imperfect purt-am/ai/a/am/aţi/au
‘carried’, which is not only obviously synthetic, but is undoubtedly the direct
descendant of the Latin Imperfectum portaba-m/s/t/mus/tis/nt and thus
akin to its Italian counterpart portav-o/i/a/amo/ate/ano. However, a
comparison of the two Romanian verb-forms reveals the complete identity of the
shapes of the final morphemes in the Imperfect and of the auxiliary in the
Perfect, which is the product of the reduction of the Romanian verb avea
< Lat. habere ‘have’.
Two conclusions follow from this coincidence. One is that the
grammatical component of a composite form can develop not only from a former
separate auxiliary word in an analytical combination, but also from a former
inflection acquiring a new status within the framework of a typologically
restructured language system. The other conclusion concerns the peculiar
relationship between an auxiliary in an analytical formation and its identical
counterpart in a grammatical compound. Since a lexical compound preserves that
status only as long as each of its elements retains its identity with the same
item as a separate word, and since there is no reason to treat grammatical
compounds differently, the grammatical part of a composite form and the
auxiliary identical with it should be regarded as one and the same language
unit serving in two different grammatical formations and thus carrying two
different categorial meanings. For instance, in the above-mentioned Romanian
example the element am/ai/a/am/aţi/au conveys the categorial
meaning of the Perfect in comparatively free pre-verbal position, while in
inseparable post-position to the verb it carries the opposite categorial
meaning of the Imperfect. Note that the ability of a language unit to express
different meanings in different positions is in full accordance with the basic
typological properties of analytical language systems.
2. GERMANIC DEVELOPMENTS
As far as the relation between the two opposing typological trends is
concerned, the Germanic group differs greatly from the two other language
groups considered above. While the trend towards analyticization has been
rather weak in the evolution of the Slavic languages, and a much stronger force
active in that direction in the Romance group has met with considerable
counteraction resulting in wide-spread re-syntheticization, the typological
evolution in the Germanic group has on the whole shown a decisive preponderance
of the analyticizing trend. The crucial testimony is neither the emergence of
new analytical forms nor the decline of inflected forms, since the former may
in time get syntheticized and thus restore the inventory of the latter. It is
much more significant that no large-scale syntheticization of analytical
formations has taken place in the typological evolution of the Germanic
languages. Only two such developments are observed in the group, and both are
in fact restricted to the Scandinavian branch.
One is the syntheticization of the Reflexive with the pronominal
auxiliary sik > sk > s~st (Icel.), which occurred at an early
stage, before the analyticizing trend started gaining ground. True, there is a
similar Western Germanic development, namely the syntheticization of the
Reflexive with the pronoun sich in spoken Yiddish, but this is obviously
a marginal process induced by a strong Slavic influence on that particular
language.
The other instance of syntheticization in the Scandinavian subgroup is a
more recent process involving the combination of a noun with the definite
article, and it has undoubtedly been facilitated by the post-nominal position
of the latter. The process has evidently reached a stage at which the article,
unlike its counterparts in all the other Germanic languages, can no longer be
regarded as an auxiliary in an analytical formation with the grammatical meaning
of noun definiteness. But the stage it has reached is not that of an
inflection, which, it must be noted, has not been attained by the Reflexive -s
either. The latter displays the properties of an agglutinate, conveying a
single categorial meaning and taking a fixed position at the end of the
syntheticized verb-form (Sw. kalla-de-s, Dan. kalde-de-s ‘was
called’). Similar properties are observed in the definite article with a single
categorial meaning and a fixed place in the synthetic noun-form between the
indicators of number and case (Sw. flick-or-na-s ‘of the girls’, Dan. by-er-ne-s
‘of the towns’).
The analytical formations that have emerged in the Scandinavian subgroup
thus display a syntheticizing trend which, however, has not proved strong enough
to bring the formations involved in the process to its conclusive inflectional
stage. But stopping at the less advanced stage of agglutination is not the only
possibility for weak syntheticization, as it can also cease after reaching the
still less advanced stage of a grammatical compound. The latter outcome is
quite probable in Danish as the most analyticized of the Scandinavian
languages.
3. A TYPOLOGICAL INNOVATION IN DANISH
An instant of grammatical composition is indeed observed in the definite
forms of Danish nouns in the Singular, e.g. sved-en ‘the sweat’, vand-et
‘the water’, which display a significant morphonological distinction from
simple formations with homophonic affixes, respectively the participle sved-en
‘scorched’ and the adjective vand-et ‘watery’. The rules for the use of
the Danish stød require its loss in a stem followed by an affix, e.g. hu's
‘house’ - huse ‘houses’, vil'd - vilde ‘wild’, skri'v
‘write!’ - skrive ‘to write’, but it is retained in the first part of a
compound. The stødless sveden, vandet with adjectival or
participial suffixes represent the former case, while the definite nouns sve'den,
van'det with the stød preserved must be regarded as grammatical
compounds.
Just as identification with an independent word is obligatory for an
element of a lexical compound, recognition of the affixed definite article in
its two gender varieties -en, -et as an element in a grammatical
compound is fully dependent on the possibility of its identification with en~et
as an auxiliary word in an analytical formation. The affixed (enclitical)
definite article and the proclitical indefinite article en~et must
consequently be regarded as two employments of the same grammatical auxiliary.
Indeed, both function within the framework of the same grammatical category of
noun definiteness, and each conveys one of its mutually opposed categorial
meanings by taking one of the two correlated positions vis-à-vis the noun. The
historic fact that the two articles have developed from quite different words has
evidently failed to prevent their eventual convergence.
But the properties of the definite article as regards the use of the
stød are shared by another grammatical element in Danish, namely -er as
the indicator of the Present in verbs, e.g. skri'ver ‘write(s)’ with the
stød preserved as opposed to the noun skriver ‘writer’ with the loss of
the stød before the agentive suffix -er. The verb-form skri'ver
must therefore be recognized as a grammatical compound, i.e. having the same
status as the definite form of the nouns. However, unlike the definite article,
which started its grammatical evolution as a demonstrative pronoun and has thus
undergone a certain measure of syntheticization, the modern indicator of the
Present was originally a verb inflection of person and number in the Present,
so that its evolution has proceeded in the opposite direction - from stronger
to weaker syntheticization.
Such convergent development is perhaps natural in a language system
consisting of two widely divergent domains, one synthetic and the other
analytical, with little to bind them into a typologically integrated coherent
ensemble. Since systems with a gap like that are unlikely to function
optimally, there must be some kind of convergence between the synthetic and
analytical parts of the language system. Such convergence may come about not
only by analytical formations being syntheticized, but also, as shown above for
the Imperfect in Romanian, by older synthetic forms moving to meet the newly
syntheticized forms halfway and entering into new systemic relations with them.
The latter are essentially relations between two employments of the same
language unit - as an auxiliary word in an analytical formation and an affixed
element in a synthetic form, conveying two different, mutually opposed meanings
of a grammatical category. Such a unit thereby acquires the typically
analytical property of its semantic contents being determined by its position.
But by acquiring positional and semantic variability the unit, when used as an affixed
element in a synthetic form, loses its original status of an inflection, which
is positionally and semantically invariable by definition.
Under the changed typological conditions the new status of the shiftable
grammatical indicator when used as an affixed element is that of second part in
a grammatical compound, intermediate between auxiliary and inflection. It
should be stressed that a radical change in the linguistic status of a language
unit need not be accompanied by a material change in its sound shape, and a
change like that may therefore escape being registered by historical
linguistics. However, a certain peculiarity in the behaviour of the unit in
question may help to reveal its newly acquired status. The peculiar treatment
of the stød in Danish verb stems followed by the indicator of the Present
testifies to the new status of the verb-form as a grammatical compound and thus
to the loss of its former status of an inflected form.
Besides the peculiar behaviour of the stød in the verb-forms of the
Present, another proof is obligatory before -er in forms like skri'ver
‘write(s)’ or ly'ser ‘shine(s)’ can be recognized as composite: it must
be shown to be in grammatical correlation, both semantically and positionally,
with a separate auxiliary word. That is evidently the link-verb er
‘am/is/are’, with which the former inflection has thus entered into a new
grammatical relationship. Semantically the latter can be described as the
opposition between two kinds of predication: processive in Solen lyser
‘The sun shines’ and qualitative in Solen er lys ‘The sun is bright’.
The parallelism should be noted between the paradigms of many Danish
nouns and verbs, containing analogous forms both with and without the stød,
i.e. simple forms with suffixes and composite forms: cf. the bare stem ly'd ‘sound’
with the stød as the Singular of the indefinite noun or as the Imperative of
the verb; the suffixed forms lyde with the stød lost as the Plural of
the indefinite noun or as the Infinitive of the verb; the composite forms ly'den
and ly'der with the stød preserved as the definite noun or as the verb
in the Present respectively.