Typically, young children’s language outputs successively approach their inputs (from their parents, peers, other caregivers and teachers), until they are identical or nearly so. But rarely, they create ephemeral Blind Alley Developments (BADs), which systematically deviate from their inputs in the course of their monolingual or simultaneous bilingual language acquisition. These developments in morphology and syntax have either no source in the target language and in their parental input (strong BADs: the examples come from Greek, Russian and Polish) or if they have one (weak BADs: the examples come from Greek, German, Croatian and French), they systematically develop their production in a wrong direction away from their parental inputs, in contrast to the typical developments where children’s outputs successively approach more and more parental inputs. In the case of simultaneous bilingual acquisition (since birth), children can produce a specific weak BAD of an overlay of the two languages’ target structures, so that the four Viennese children that are observed, form German noun plurals according to Croatian conditions and Croatian noun plurals according to German conditions. For their analysis, the conception of superposition in quantum physics can be used that Mattiello and Dressler ^[1] had already adapted to linguistics. Nobody else before has detected, analysed and explained these phenomena. Whereas no acquisition models other than the Natural Linguistic model can explain strong morphological BADs, except with fundamental changes in their models; weak morphological BADs and the Polish strong syntactic BAD discussed could be explained by other models, if they adopt argumentations described in this contribution.

BADs represent the best evidence that children do not simply imitate their parents, etc., but self-organise their language acquisition (autopoiesis). The explanation of such radical types of self-organisation may be considered to fall into the trap of Feyerabend’s principle of “Anything goes!”, which renders decisions between competing explanations impossible or difficult. We think that we have overcome this danger.

Another type of self-organisation occurs much earlier in the premorphological period, when children rather unsystematically create few examples of extragrammatical morphological patterns, especially reduplications, blends and onomatopoetic forms such as echo words. These unsystematic developments such as G. mapi, a blend of Mama ‘mumj’ and Pap+i ‘dadd+y’ or the hypocoristic Lis+i (which has become permanent for my elder daughter) from Elisabeth, are not discussed in this entry, which exclusively analyses and explains the systematic, albeit ephemeral developments of BADs.

This premorphological period is followed by the protomorphological period, when children start to acquire adult structures, until they acquire the core of adult morphology (cf. ^[2][3]).

The linguistic model espoused in this entry is that of Natural Linguistics ^[4][5]: it is a semiotic and cognitive model of universal parametrized preferences, e.g., for iconic relations between form and meaning and for expressing marked morphological and syntactic structures by more salience than their unmarked counterparts. Children can follow a universal preference in a production, even if they do not find an example of it in their input. In contrast, input-based acquisition models (incl. Constructional Morphology and Syntax) postulate that children can produce only what they have found in their language input. Generative nativist models rely for child acquisition on specific (not semiotically- or cognitively-based) linguistic parameters of Universal Grammar. Since cognitive and semiotic bases must be assumed for children’s maturation, this contradicts Occam’s Razor. Moreover, no respective parameters have been assumed so far for the Blind Alley Developments discussed in this contribution.

The language acquisition model on which this entry is based not only fits to the above characterised linguistic model, but conforms to the universally accepted methodology of MacWhinney ^[6] with specifications described in the appendix on methods, plus Popperian falsificationism in the interpretation of unclear cases according to ^[7].