Poznań Studies in Contemporary Linguistics 45(1), 2009, pp.

33–42
© School of English, Adam Mickiewicz University, Poznań, Poland
doi:10.2478/v10010-009-0003-9

NATURAL PHONOLOGY AS PART OF NATURAL LINGUISTICS

WOLFGANG U. DRESSLER
Vienna University and Austrian Academy of Sciences
[email protected]

ABSTRACT

In the history of phonology, no model of phonology (e.g. Trubetzkoy’s model or Optimality The-
ory) has stayed for any considerable time alone on its own, i.e. without parallel models of syntax
and morphology. Thus, Stampean Natural Phonology has been complemented in Europe by Natu-
ral Morphology, Natural Syntax and Natural Text Linguistics. In this way, Natural Phonology can
be integrated into an overall model of Natural Linguistics. This model has been further unified by
giving it the status of a preference theory and by providing functionalist epistemology and a se-
miotic metatheory for all of its parts, including Natural Phonology. This unity of Natural Linguis-
tics is demonstrated with the universal preference parameters of figure vs. ground sharpening,
binarity, iconicity, indexicality and the naturalness scale biuniqueness > uniqueness > ambiguity.
The final argumentation focuses on the “weak claim” of universality of preferences (against the
“strong claim” of innate phonological processes) by understanding universal preferences (includ-
ing phonological processes) as plausible constructivist “software” strategies which are con-
structed in language acquisition for overcoming universal innate “hardware” problems.

KEYWORDS: Natural Phonology; Natural Linguistics; semiotics; preference theory; weak con-
structivist claim on universality.

The pioneering work on Natural Phonology (henceforth NPh) by David Stampe (1969,
1979) has treated phonology on its own, relating it to phonetics, introducing a function-
alist framework and insisting on the importance of external or substantive evidence. He,
and later his research associates (Donegan and Stampe 1979, cf. Hurch and Rhodes
1996), have from the beginning introduced and then elaborated a model of natural pho-
nological processes and opposed them to learned rules of morphology and morphonol-
ogy. But these two areas were characterized just negatively.
The late 1970s, and especially the 1980s, have seen the rise of the sister disciplines
of Natural Morphology (henceforth NM; Mayerthaler 1981; Wurzel 1984; Dressler et
al. 1987; Kilani-Schoch and Dressler 2005), Natural Textlinguistics (henceforth NT:
34 W. Dressler

Dressler 1989, 1996b; Merlini Barbaresi 1988), and Natural Syntax (henceforth NS:
Mayerthaler and Fliedl 1993; Mayerthaler et al. 1998; Orešnik 2004), and also of the in-
terfaces of phonology and morphology in morphonology (Dressler 1985, 1996c) and of
syntax and morphology in morphosyntax (Mayerthaler et al. 1998; Orešnik 2004).
These developments have opened the horizon of treating all these sister disciplines in a
similar way and with similar concepts, and of considering them as parts of Natural Lin-
guistics. A workshop at the Societas Linguistic Europaea meeting in Kraków organized
in 1996 by Katarzyna Dziubalska-Kołaczyk had this perspective explicitly (cf. a similar
a workshop at the SLE meeting in Forlì in September 2008). This may be compared to
the rise of Optimality Theory in phonology (influenced by NPh), and its extension to
other areas of grammar (cf. Prince and Smolensky 1993; de Lacy 2006).
Thus, already in the late 1970s, I was against continuing to work on NPh “in splen-
did isolation”, as if phonological naturalness had nothing in common with morphologi-
cal, syntactic and textual naturalness. Rather, I felt it was imperative to study the com-
monalities and differences between various types of naturalness, and to devise both a
consistent metatheory and compatible methodologies, if possible, for all or most of
them (Dressler 1984, 1985, 1996a, b). Otherwise, we cannot systematically account for
synchronic and diachronic interactions between language components and stand less
chance of competing successfully with models which encompass all or many compo-
nents of language.
Let me start with the bases of naturalness. Classical NPh concentrated on the pho-
netic bases of phonology. However, if cognitive, psychological and sociopragmatic
bases play a role in NM, NS, and NT, then it is highly unlikely that they were negligible
in NPh. Some important non-phonetic factors motivating aspects of phonological natu-
ralness are as follows:

(a) The rhythmic organization of sequential activity is not only a property of

speech production and perception and music but of any motor activity, thus
rhythmicity in prosodic phonology is just a special case (cf. already Mayer-
thaler 1982: 221f; Auer 1990: 14).

(b) Now, if we look into semiotics, there is the related principle of figure and
ground (cf. Dressler 1990a: 81–2, 91–2; Kilani-Schoch and Dressler 2005)
which predicts that figures tend to be foregrounded, grounds to be further
backgrounded. As a result, within a contrasting pair or triple, the perceptually
more salient partner tends to be further enhanced in perceptual saliency, the less
salient partner(s) tend to be further backgrounded. Therefore, the well-known
“rich-get-richer” principle of NPh, introduced by Patricia Donegan (1978: 143),
may be subsumed under the more general semiotic principle of figure and
ground, insofar as the phenomenon that comparatively “weak” elements are
more liable to be further weakened rather than strengthened, whereas compara-
tively “strong” elements are more likely to be strengthened than weakened, can
 5atural Phonology as part of 5atural Linguistics 35

be seen as a phonological consequence of the tendency towards contrast sharp-

ening between figure and ground. This may be assumed on two conditions:
firstly, prosodically strong positions are identified as figures, prosodically weak
positions as grounds; secondly, the rich-get-richer principle must be con-
strained to account for strengthening in strong positions and for weakening in
weak positions. Thus the rich-get-richer principle ceases to be an ad hoc princi-
ple, because it can be subsumed under a semiotic principle which holds for per-
ception and the perceptive basis of language in general.

(c) There is a universal preference for binary paradigmatic and syntagmatic con-
trasts (cf. Dressler 1990a: 85; Kilani-Schoch and Dressler 2005). It is both
semiotically based (note Peirce 1965: II.377), and it seems to be neurologically
based (starting with the binary choice between presence and absence of neural
firing) and holds both for phonology and for other linguistic components. For
phonology, note the preference for binary distinctive features, the opposition
between culminative vs. demarcative prosodic signals, and other phonological
antagonisms, cf. already Vincke (1989). For morphology, note e.g. the binary
contrast between head vs. non-head, for syntax and text, but also for phonology
the preference for binary not n-ary figure vs. ground contrasts.

(d) A last and related example: There is an interplay and binary contrast between
fortition processes, which I have renamed foregrounding processes (cf. also
Luschützky 1988; Dogil and Luschützky 1990) and lenition processes, which I
have renamed backgrounding processes. This interplay is based not only in
phonetics and in general rhythmic behaviour (cf. (a) above), but also in the so-
ciopragmatic foundation of communication in the interaction between speaker
and hearer (cf. Dressler and Wodak 1982; Dressler and Moosmüller 1991), i.e.
rhythmic alternation and sociopragmatics are not only secondary intervening
variables in actualization of speech behaviour (and occasional sources of un-
naturalness) but a conditio sine qua non of the antagonism between foreground-
ing and backgrounding (cf. also Auer 1990: 16). Therefore, Dziubalska-
Kołaczyk (2002) spoke of speaker-friendly vs. hearer-friendly processes.

Let us illustrate these concepts with examples from Polish and Czech prosodic phonol-
ogy and its consequences (cf. Madelska and Dressler 1996): the preference for figure-
and-ground sharpening results in the prosodic preference for demarcative accentuation,
i.e. for having most prosodic prominence at the beginning or end of a word. This is
based in processing according to the so-called bathtub effect, i.e. the beginning and end
of a word are the most salient parts. Young children may even reduce words to the very
beginning and the end; note the following Italian diminutives produced by two-year old
children: t-ina ← adult tartarugh-ina ‘turtle-DIM’, m-ina ← man-ina ‘hand-DIM’
(Noccetti et al. 2007).
36 W. Dressler

Czech prosody is characterized by an obligatory initial accent, as in Slovak, Hun-

garian, etc., as in e.g. sobota (sww) ‘Saturday’, or the French loan word niveau. The as-
signment of secondary accents follows the preference for binarity, and therefore, for bi-
nary feet and for falling accent, e.g. exemplifikace (swswsw) ‘exemplification’, uni-
verzitní (swwsw) ‘universitarian’.
Polish combines the preferences for final accent, falling accent and binary feet in
preferring word-final trochees as the seat of the main accent, word beginning as the seat
of secondary accents, as in operowy (swsw) ‘operatic’, uniwersytecki (swswsw) ‘uni-
versitarian’. In substandard Polish, even loan-words with antepenultimate accent taken
over from Latin adopt penultimate main accent in order to follow the preference for
word-final trochees, as in: logika ‘logic’ sww → wsw, uniwersytet ‘university’ swsww
→ swwsw.
When developing metatheoretic bases for NM and NT, I took the same approach to
NPh, too. This had to do, first, with functionalist epistemology, for which I would like
refer to my publications (cited in Dressler 1985, 2002). What I want to expose here are
the semiotic foundations for NPh I have tried to develop. If we stick to the sign charac-
ter of language, and if we want to compare language components among themselves,
and language as a whole with other sign systems, such as music or non-verbal commu-
nication, we need an appropriate meta-theory. Semiotics seems to be the best choice,
particularly Peircean semiotics, as argued, for example, in Dressler (1989a, 1990a: 76–
78, 1996a–c). Semiotics can be construed, in equivalent ways, either as a relational the-
ory, or as an action theory, so that phonological processes as actions can be subsumed
under human actions in general.
As already described in Dressler (1984, 1985: Chapter 10), several semiotic pa-
rameters are of value not only in NM, NS, and NT, but also in NPh. If we deduce, from
semiotics and other bases of phonology, universal preferences for each of these semiotic
parameters, and assume that, ceteris paribus, each preferred option thus deduced is also
actually preferred over corresponding dispreferred options, we can explain the cross-
linguistic frequency distributions which I am going to discuss below.
But first let us refer to Peircean semiotics (cf. Peirce 1965). Among Peirce’s
elementary triad of signs (icon, index, symbol), an icon is a sign “which exhibits a similar-
ity or analogy to the subject of discourse” (Peirce 1965: I.369). A more iconic sign is more
natural than a less iconic sign. This holds also for phonological processes (henceforth PRs)
as signs. Following this preference for more iconicity, PRs, as natural phonological proc-
esses, are more iconic, and therefore also more frequent, than MPRs (learned morpho-
nological rules). For example, Polish palatalisation has a more iconic relation between in-
put and output when it is a PR rather than an MPR. Compare the Polish PR of surface
palatalisation k → k’ / _ i, j (as in the pejorative plural Polak-i from Sg. Polak) with the
corresponding MPR k → c/ _ y in the normal plural Polac-y. And among PRs, intrinsic al-
lophonic PRs are the most iconic PRs, and therefore also the most frequent PRs. On the
other hand, deletion processes are rare outside fast/casual speech, because a zero output of
a non-zero input is non-iconic. Finally, PRs of metathesis are rare because they are anti-
 5atural Phonology as part of 5atural Linguistics 37

iconic: e.g. the Ancient Greek root /tek/ ‘give birth’ is iconically preserved in the Aorist é-
tek-en ‘she gave birth’, whereas the zero grade of the root /tk/ in 1. Sg. Pres. /tí-tk-ō/ is
obligatorily changed to [tíktoː], where the phonologically illegal order of consonants in the
input contradicts the legal order of consonants in the output.
Let us pass to the second non-conventional Peircean sign type: An index is a sign
“which like a pronoun demonstrative or relative, forces the attention to the particular object
intended (= signatum) without describing it” (Peirce 1965: II.369). An indexical sign is the
easier to grasp, the closer the signans and signatum are. From this adjacency preference,
we can deduce the following phonological indexicality preferences:

(a) context-sensitive PRs are indexical and, therefore, predictably also more fre-
quent than context-free PRs, which are non-indexical;

(b) PRs of contact assimilation are more indexical and thus predictably also more
frequent than PRs of distant assimilation.

Context-sensitive PRs are phonologically indexical insofar as they refer to their relevant
phonological context. MPRs, however, are both morphologically and phonologically
indexical. For example, the English MPR of velar softening in elektri/k/ → electric-ity
refers both to the morphological context of certain suffixes where it applies, and to the
phonological context of a following /i/. Since, semiotically, morphology comes before,
and is more important than, phonology, morphological indexicality is more important
than phonological indexicality. This semiotic precedence explains the unidirectional
change from PRs to MPRs in diachrony, and the possible loss of phonological indexi-
cality in MPRs, which then turn into allomorphic rules, whereas the loss of morpho-
logical indexicality and the preservation of phonological indexicality of an MPR is un-
heard of.
The reliability of a sign (Morris 1938) is highest in the case of a biunique relation
between its signans and signatum – in phonology, between input and output of a PR.
Thus, biunique PRs are more natural than unique PRs, and these in turn more natural than
ambiguous PRs. MPRs are never biunique but always neutralising (therefore at best
unique). If there is a biunique relation, the productivity of the respective process is guaran-
teed, but a neutralising unique process may also be productive. If a phoneme is always
realised by the same intrinsic allophone, and if this allophone can never be derived from
another phoneme, then this PR is biunique. Thus the semiotic preference for (bi)
uniqueness helps to explain, in convergence with iconicity predictions, (1) why
processes are general and productive and (2) why intrinsic allophonic PRs are by far the
most frequent PRs, and why allophonic processes are more frequent than phonemic
processes
NPh has
or MPRs.
been a pioneer in systematically using “external” or substantive evidence
on a par with internal evidence (cf. Churma 1985; Singh 1988). But the validity of “ex-
ternal evidence” hinges on the theoretical underpinnings of a bridge theory integrating
phonological theory and the respective theory underlying empirical investigations (cf.
38 W. Dressler

Botha 1984). Thus, I have tried to contribute at least explanatory sketches of bridge
theories for sociophonology (Dressler and Wodak 1982; Dressler and Moosmüller
1991) by integrating NPh with sociolinguistic and psycholinguistic models, or to a re-
lated scenario of diachronic change (cf. Dressler 1984, 2003) by integrating theory of
naturalism with models of diachronic change.
Together with phonetic evidence, the most important substantial evidence for NPh
has come from language acquisition, particularly first language acquisition of phono-
logy. Here, as for NM and NT, we have combined naturalness theory with the construc-
tivist model of self-organising processes, which provides a bridge theory for the weak
claim on innateness of phonological processes (cf. Dressler 1984: 30, 1988: 15) as op-
posed to Stampe’s (1969 etc.) strong claim (expressed more cautiously Donegan 1985:
26, with note 5; cf. discussion in Hurch 1988a; Dziubalska-Kołaczyk 2002).
Stampe (1969) claims that “the phonological system of a language is largely the
residue of an innate system of phonological processes, revised in certain ways by lin-
guistic experience”. In contrast, the weak claim restricts this hypothesis on first (and,
partially, also natural second) language acquisition to: “children have to systematically
inhibit the process types which are (or become by maturation) available to them”. Both
claims acknowledge the phonetic basis of phonological processes, but (in his publica-
tions) Stampe equates the set of universal processes with the language-innocent state of
phonology before children start to restrict, order or suppress universal processes, cf.
Stampe (1980: ix): “in its language-innocent state, the innate phonological system ex-
presses the full system of restrictions of speech”.
There are important empirical problems with such a strong claim (cf. Dressler 1984:
47; 1985: 210f, 252f). Especially natural processes observed in small children are to-
tally absent first, or initially appear only in a very limited and even irregular way
(which resembles casual speech processes or even MPRs). Examples are provided by
phonological (“surface”) palatalisation in Polish or final obstruent devoicing in many
languages. Thus, Smoczyński (1955), in his very thorough description, observes that his
two children Anka and Pawełek initially did not palatalize velars and other consonants
before /i/ and the glide /j/ – i.e. the most palatal vowel and glide – although this is an
obligatory natural process in adult Polish. So why should Polish children suppress an al-
legedly inherited universal natural process, if this very process is omnipresent in their
input? The children only started applying the process later, and then in an exceptionless
way. Similarly, Locke (1983) reports on a South African boy who first did not apply the
natural Afrikaans process of final devoicing to word-final obstruents. Moreover, any
similar strong claim would be nonsensical in NM, NS and NT.
In contrast, the weak claim is that (substance-based) universal preferences, when
available to a child at a given stage of extralinguistic and linguistic maturation, can be
taken up by the child even when contradicted by language-specific facts equally avail-
able to the child. This claim is compatible, on all levels, with both natural theories and
acquisition evidence. Thus, universal phonological processes are not innate themselves
but universally likely reactions to phonetic (and other phonology-related) difficulties to
 5atural Phonology as part of 5atural Linguistics 39

be overcome; for example, final devoicing is a plausible reaction to the phonetic diffi-
culty of producing word-final voiced obstruents before a pause.
The concept of self-organising processes in language acquisition (see Karpf 1990)
assumes an interplay of genetic preprogramming and of selecting and evaluating post-
natal information within and among preferentially coupled neuronal assemblies (which
develop into interacting modular systems). In contrast to Chomskyan innativism, much
less genetic preprogramming is assumed, and not necessarily of an autonomously
grammatical nature. Thus, a radical constructivist view would be to assume for phonol-
ogy only preprogramming of phonetic processors and of general cognitive principles,
and the emergence of (non-innate!) phonology as the outcome of the organisation and
reorganisation of processing phonetic information.
In this way, Cavalcante Albano (1989) has interpreted the genesis of deviant phono-
logical forms in small Portuguese/Brazilian children such as transient andora following
earlier andó corresponding to adult andou ‘has walked’; the final -a is claimed to have
originated from a deviant interpretation of a schwa-like off-glide often following word-
final –r, and in reaction to the child’s own -r-dropping. Such an analysis is fully com-
patible with the NPh concepts of applying a universal phonological process, such as the
weakening and centralisation of [a], to the perception and interpretation of a low schwa,
and of “rolling back” the natural process of -r-dropping.
However, the self-organising processes model predicts, first of all, that such phono-
logical processes are not used (either in production or perception or evaluation) all at
once, but arise at different times of maturation. Second, it does not predict that a child’s
phonological processes must be absolute constraints on production and perception, in-
sofar as alternative organisations and reorganisations of processing are possible. Both
predictions fit actual data of language acquisition better than the above-mentioned
strong claim of classical NPh.
Since the phonetic information to be handled becomes always more complex during
maturation, many phonological processes are likely to arise or be generalized only at
later points of maturation. And this is fully compatible with what was found already by
Seifert (1985) and Karpf (1990) about the acquisition of morphology with respect to the
claims of NM. In maturation, “phonology grows”, and not just language-specific pho-
nology, but also universal parts of phonology.
Thus, the above-mentioned weak claim on phonology acquisition fits the data better
and is in better alignment with the facts of acquisition of morphology and syntax (cf.
Karpf 1990). It can be integrated into a bridge theory of language acquisition which re-
lates physiology, psychology, and neurology to NPh (cf. Dressler 1997, 2003). More-
over, as acknowledged by Donegan (1985: 30, note 5) “it would not alter the theory of
natural phonology substantially to say that processes may be discovered by the child as
he learns to use his vocal tract”; in fact, this brings NPh closer to NM, to NS and NT.
In conclusion, we may state that integrating NPh, together with NM, NS and NT,
into a general theory of naturalness allows several theoretical, methodological and em-
pirical advantages.
40 W. Dressler

REFERENCES

Auer, P. 1990. “A note on prosody in Natural Phonology”. In: Méndez Dosuna, J. and C. Pen-
sado (eds.), 5aturalists at Krems. Salamanca: Ediciones Universidad de Salamanca. 11–22.
Botha, R. 1984. “A Galilean analysis of Afrikaans reduplication”. Stellenbosch Papers in Lin-
guistics 13.
Cavalcante A.E. 1989. “Self-organizing processes in the acquisition of phonology”. Revue de
Phonétique Appliquée 91–93. 131–145.
Churma, D.G. 1985. Arguments from external evidence in phonology. New York: Garland.
De Lacy, P. 2006. Reduction and preservation in phonology. Cambridge: Cambridge University
Press.
Dogil, G. and H.C. Luschützky. 1991. “Notes on sonority and segmental strength”. Rivista di
Linguistica 2. 3–54.
Donegan, P.J. 1978. “On the natural phonology of vowels”. Working Papers in Linguistics, The
Ohio State University 23.
Donegan, P.J. 1985. “How learnable is phonology?” In: Dressler, W. and L. Tonelli (eds.), Pa-
pers on natural phonology from Eisenstadt. Padova, Italy: CLESP (Cooperativa Libraria
Editoriale Studentesca Patavina). 19–31.
Donegan, P.J. and D. Stampe. 1979. “The study of Natural Phonology”. In: Dinnsen, D. (ed.),
Current approaches to phonological theory. Bloomington: Indiana University Press. 126–
173.
Dressler, W.U. 1984. “Explaining Natural Phonology”. Phonology Yearbook 1. 29–51.
Dressler, W.U. 1985. Morphonology. Ann Arbor: Karoma Press.
Dressler, W.U. 1988. „A linguistic classification of phonological paraphasias”. In: Dressler,
W.U. and J. Stark (eds.), Linguistic analyses of aphasic language. New York: Springer. 1–
22.
Dressler, W.U. 1989a. Semiotische Parameter einer textlinguistischen 5atürlichkeitstheorie.
Wien: Verlag der Österreichischen Akademie der Wissenschaften.
Dressler, W.U. 1989b “Markedness and naturalness in phonology: The case of natural phonol-
ogy”. In: Tomic, M. (ed.), Markedness in synchrony and diachrony. Berlin: de Gruyter.
109–119
Dressler, W.U. 1990a. “The cognitive perspective of ‘naturalist’ linguistic models”. Cognitive
Linguistics 1. 75–98.
Dressler, W.U. 1990b. “Marked and unmarked translation: An approach from semiotically based
Natural Text Linguistics”. Meta 35. 138–148.
Dressler, W.U. 1996a. “Principles of naturalness in phonology and across components”. In:
Hurch, B. and R. Rhodes (eds.). 41–51.
Dressler, W.U. 1996b. “Parallelisms between Natural Textlinguistics and other components of
Natural Linguistics”. Sprachtypologie und Universalien-Forschung 49. 295–311.
Dressler, W.U. 1996c. “A functionalist semiotic model of morphonology”. In: Singh, R. (ed.),
Trubetzkoy’s orphan. Amsterdam: Benjamins. 67–83.
Dressler, W.U. 1997. “‘Scenario’ as a concept for the functional explanation of language
change”. In: Gvozdanović, J. (ed.), Language change and functional explanations. Berlin:
Mouton de Gruyter. 109–142.
Dressler, W.U. 2002. “Naturalness and functionalism”. In: Dziubalska-Kołaczyk, K. and J.
Weckwerth (eds.), Future challenges for 5atural Linguistics. München: Lincom Europa.
83–101.
 5atural Phonology as part of 5atural Linguistics 41

Dressler, W.U. 2003. “Naturalness and morphological change”. In: Joseph, B.D. and R.D. Janda
(eds.), The handbook of historical linguistics. Oxford: Blackwell. 461–471.
Dressler, W.U., W. Mayerthaler, W. Wurzel and O. Panagl. 1987. Leitmotifs in 5atural Mor-
phology. Amsterdam: Benjamins.
Dressler, W.U. and S. Moosmüller. 1991. “Phonetics and phonology: A sociopsycholinguistic
framework”. Phonetica 48. 135–148.
Dressler, W.U. and R. Wodak. 1982. “Sociophonological methods in the study of sociolinguistic
variation in Viennese German”. Language in Society 11. 339–370.
Dziubalska-Kołaczyk, K. 2002. “Challenges for Natural Linguistics in the twenty-first century:
A personal view”. In: Dziubalska-Kołaczyk, K. and J. Weckwerth (eds.), Future challenges
for 5atural Linguistics. München: Lincom Europa. 103–128.
Hurch, B. 1988. “On the notion ‘latent process’”. In: Bertinetto, P.M. and M. Loporcaro (eds.),
Certamen Phonologicum I. Torino: Rosenberg & Sellier. 103–115.
Hurch, B. and R.A. Rhodes (eds.). 1996. 5atural Phonology: The state of the art. Berlin: Mou-
ton de Gruyter.
Karpf, A. 1990. “Selbstorganisationsprozesse in der sprachlichen Ontogenese: Erst- und Fremd-
sprache(n)”. Tübingen: Narr.
Kilani-Schoch, M. and W.U. Dressler. 2005. Morphologie naturelle et flexion du verbe français.
Tübingen: Narr.
Locke, J. 1983. Phonological acquisition and change. New York: Academic Press.
Luschützky, H.C. 1988. “Sixteen possible types of natural phonological processes”. Wiener Lin-
guistische Gazette 42/43. 79–103.
Madelska, L. and W.U. Dressler. 1996. “Postlexical stress processes and their segmental conse-
quences illustrated in Polish and Czech”. In: Hurch, B and R. Rhodes (eds.). 189–200.
Mayerthaler, W. 1981. Morphologische 5atürlichkeit. Wiesbaden: Athenaion.
Mayerthaler, W. 1982. “Markiertheit in der Phonologie”. In: Vennemann, T. (ed.), Silben, Seg-
mente, Akzente. Tübingen: Niemeyer. 205–246.
Mayerthaler, W. and G. Fliedl. 1993. “Natürlichkeitstheoretische Syntax”. In: Jacobs, J., A. von
Stechow, W. Sternefeld and T. Vennemann (eds.), Syntax. Berlin: de Gruyter. 610–635.
Mayerthaler, W., G. Fliedl and C. Winkler. 1998. Lexikon der natürlichkeitstheoretischen Syntax
und Morphosyntax. Tübingen: Stauffenburg.
Merlini Barbaresi, L. 1988. Markedness in English Discourse: A semiotic approach. Parma:
Zara.
Morris, C.W. 1938. Foundations of the theory of signs. Chicago: University of Chicago Press.
Noccetti, S., A. de Marco, L. Tonelli and W.U. Dressler. 2007. “The role of diminutives in the
acquisition of Italian morphology”. In: Savickienė, I. and W.U. Dressler (eds.), The acquisi-
tion of diminutives. Amsterdam: Benjamins. 125–153.
Orešnik, J. 2004. 5aturalness in (morpho)syntax: English examples. Ljubljana: Slovenska
Akademija Znanosti in Umetnosti.
Peirce, C.S. 1965. Collected papers. (Ed. by C. Hartshorne and P. Weiss.) Cambridge, MA: Har-
vard University Press.
Prince, A. and P. Smolensky. 1993. Optimality Theory. Constraint interaction in generative
grammar. Cambridge, MA: MIT Press.
Seifert, K. 1985. Ikonizität von Pluralformen. Vienna: Dissertationen der Universität Wien.
Singh, R. 1988. “Linguistic theory and external evidence”. Canadian Journal of Linguistics
33(4). 329–511.
Smoczyński, P. 1955. Przyswajanie przez dziecko podstaw systemu językowego [The acquisition
of the fundamentals of the linguistic system by children]. Łódź: Ossolineum.
42 W. Dressler

Stampe, D. 1969. “The acquisition of phonetic representation”. In: Binnick, R. et al. (eds.), Pa-
pers from the fifth regional meeting of the Chicago Linguistic Society. Chicago: Chicago
Linguistic Society. 443–454.
Stampe, D. 1979. A dissertation on natural phonology. New York: Garland.
Vincke, J.L. 1989. “Binarism vs. multinarism: A theoretical choice or a natural hierarchy”. Uni-
versidade Federal de Santa Catarina, Working Papers in Linguistics 5. 70–74.
Wurzel, W.U. 1984. Flexionsmorphologie und 5atürlichkeit. Berlin: Akademie-Verlag.

Address correspondence to:

Wolfgang U. Dressler
Institut für Sprachwissenschaft
University of Vienna
Berggasse 11
1090 Vienna
Austria
[email protected]