The Critical Period Hypothesis

Table of contents

1. Introduction

2. The Critical Period Hypothesis
2.1 About critical periods
2.2 Assuming a critical period for human language

3. Two cases that seem to support the CPH
3.1 Victor
3.2. Genie

4. Two cases that seem to question the CPH
4.1 Isabelle
4.2 P.M. and J.M - The Czech twins

5. Discussion

6. Conclusion

7. Bibliography

1. Introduction

Encounters with feral children have been attracting intellectual examination and curiosity since the early seventeenth century. Among others, Anthropologists, Psychologists and Linguists hoped to find answers to central questions of mankind: What makes us human beings? What distinguishes us from animals? Nature or nurture, what has greater impact on us? Do we possess innate ideas or do we need society to become human? As feral children seemed to promise answers to such existential concerns, they “have had a privileged role as objects of knowledge in Western human science” (Benzaquén 2002: 65).

The concept of the human subdivision homo ferus was brought up by Carl Linnè (also known as Linnaeus) in 1758. It was initially used to describe individuals who had lived in the wilderness on their own or maybe with animals. The anthropologist Robert Zingg (1940) used the translated version feral man in his account of “extreme cases of isolation”. Shortly afterwards, the term was especially associated with children and recently also children who experienced isolation and deprivation are included in that category. So the “wildness” nowadays rather signifies a kind of strangeness or development in abnormal circumstances in general (cp. Benzaquén 2002: 66) and as such it is used in this term paper.

One field of investigation in which these case studies are frequently cited, is that of linguistics. Especially with respect to the assumption of maturational constraints on language acquisition, scholars tried to draw conclusions from the success or failure feral children exhibited after discovery. Respectively, they were interpreted as evidence or counter-evidence for the Critical Period Hypothesis (CPH) which postulates that language can only be acquired normally up to a certain age. Eric H. Lenneberg is closely connected to this assumption and it is remarkable that he did not included cases of feral children. He explicitly comments on that when he points out that such cases should not be taken as evidence because “the nature of the social and physical environment is never clear and the possibility of genetic deficiencies or congenital abnormalities can never be ruled out” (1967: 141).

Arguing in a similar fashion, Wayne Dennis (1941: 432) suggests “that these cases not be cited as evidence for any social or psychological theory”. Nevertheless, they have been taken into account in the context of the CPH because many scholars think that they offer valuable insights into the topic. This term paper attempts to clarify up to which extend they really do so.

To achieve this task, the paper is organised as follows. In Section 2 a general de- scription of critical periods is given before turning to Lenneberg’s hypothesis. Section 3 focuses on two cases that are often taken as evidence for the CPH, namely Victor and Genie. Both didn’t master language acquisition to a satisfactory level. Section 4, on the other hand, deals with three children who are expounded as counter-evidence for the CPH because they caught up quickly on language learning close to age seven. In both sections, after a short description of the individual experiences prior to dis- covery, a detailed account of linguistic achievements and shortcomings is presented. Section 5 will discuss the outcomes and put them into a broader scientific context by adding results of further research. Section 6 offers a preliminary conclusion, namely that cases of feral children should be included as indirect evidence but that this needs to be done carefully.

2. The Critical Period Hypothesis

2.1 About critical periods

Originally the term critical period emerged in the field of experimental embryology. Here it was used to describe time spans in which normal embryonic development (e.g. of minnow eggs) could be inhibited or even mutated (Stockard 1978 presented in Herschensohn 2007: 7). Later on, the term also occurred in the area of animal behaviour. A famous example is Konrad Lorenz’s experiment with a greylag gosling. He demonstrated that the young goose attached to the first large, moving object being around. This so-called imprinting occurs in the first week after hatching and turned out to be irreversible, i.e. the gosling always preferred a human, the original imprint-target, to another bird (cp. Lorenz 1937: 264). There are other interesting examples for the study of critical periods in animals including, for instance, the development of vision in kitten, the learning of birdsong or the establishment of dialectic communication systems in orca whales. Yet, for the purpose of this paper it is more significant to define critical periods as such. A single clear-cut definition, however, is not available because the term has not been used consistently in the scientific literature.

With respect to etymology, the expression “critical” in this sense derives from the Greek words krino (to decide) and krisis, meaning “a decisive turning-point in any matter” (The Encyclopaedic Dictionary 1889: 570). In some way, during critical periods indeed decisions are made and they cannot be repeated or changed afterwards. Corresponding to that, Scott (1978) gives the following definition for them:

The time of most rapid organization in a particular process is not only a critical period for decisions but is also an optimal period for producing desirable (or at least desired) changes in organization. Change cannot be produced before the organizational process begins nor after it has ceased. (Scott 1978: 359)

In other words, there occurs a phase of high susceptibility to certain stimuli which in return produces changes in the living system. Before and after this phase no or at least fewer susceptibility is given. These phases coincide with a developmental stage of the organism, i.e. with a certain age. Not in all, but in most cases, such a “peak period of plasticity” can be observed during stages of early development (cp. Newport et al. 2001: 482).

At first glance it might occur unpractical and risky that animals, including humans, are equipped with mechanisms which rely on such a restricted timing. According to Pinker (1994: 293) this impression arises because “most of us have an incorrect understanding of the biology of organisms’ life histories”. With reference to the example of metamorphosis in insects, he emphasises that “designs hang around during the times of life that they are useful, not before or after” (ibid.:294). Furthermore, Bruer (1999: 107) regards critical periods as “adaptive, clever, and efficient feats of biological engineering”. For him, they “make evolutionary sense because they rely on stimuli that are ubiquitous within normal human environments” (ibid.: 110) and the same holds for animalistic environments. During critical periods, our brain seems to expect certain input and also needs it to fine-tune its neural circuits, a fact that is called experience-expectant brain plasticity (cp. Greenough quoted in Bruer 1999: 108). Abnormal circumstances, like deprivation, therefore make a normal development impossible because the required input does not occur.

To characterise critical periods per se, scientists hold on to some parameters that have been declared to be structurally representative for them. According to Bornstein (1987: 5), there are five of these factors¹ and the following table displays them with the corresponding examples from birdsong learning (content of this table cp. Herschensohn 2007: 11):

Abbildung in dieser Leseprobe nicht enthalten

Another term that appears often in this context is “sensitive period”. While critical periods were supposed to end abruptly, these periods were thought to end gradually. This distinction turned out to be useless because scientists discovered that most critical periods end gradually (cp. Bruer 1999: 104; Newport et al. 2001: 482). Nowadays both terms are used quite interchangeably.

2.2 Assuming a critical period for human language

The ultimate strength and appeal of the assumption that there must be an age sensitivity for language acquisition lies in its link to everyday experience: While observing a young child learning his/her native tongue with no apparent effort, without special grammatical instructions and in a short period of time, probably everyone starts wondering why learning a second language later in life is that much harder. Adults are able to master a second language up to a native-like level, but only as a result of a long-lasting and labour-intensive learning process. The perfect pro nunciation of a foreign language, however, is hardly ever reached. A famous example is the German-born Henry Kissinger: he retained his accent while his younger brother does not have one (cp. Pinker 1994: 291). A similar discrepancy can be seen in the case of bilingualism. For example, in immigrant families the children become native speakers of the second language while their parents will keep their foreign accents and will make more grammatical mistakes (cp. Herschensohn 2007: 2).

All in all, it seems to be a kind of folk wisdom that we (or more accurately our brains) are equipped with the necessary tools to learn our native tongue naturally at a young age, but that this ability gets lost as we grow older. At first sight that seems to be a well-understandable and logical presumption. Nonetheless, in many scientific fields (as in linguistics, neurology and biology, just to name a few) this idea of a connecting between language and maturation has been hotly debated ever since it was brought up.

Back in the 1950s this matter was not an important issue yet. As behaviourism and structuralism were dominant perspectives, the explanation for the folk belief presented above had been: language is the outcome of mechanisms like conditioning, association, practice in exercising skills and reinforcement. As B.F. Skinner puts it, “verbal behavior” is acquired by a child “when relatively unpatterned vocalizations, selectively reinforced, gradually assume forms which produce appropriate consequences in a given verbal community” (1957:31).

The dominance of the behaviourist approach declined as experimental work headed into another direction and as perspectives shifted to a biological understanding of human language. The notion of localization, for example, had a revival and the notion of a “Language Acquisition Device” came up (cp. Strozer 1994: 135). Eric H. Lenneberg, probably inspired by these tendencies, examined language as a natural phenomenon, assuming that it can be studied in a similar fashion like other aspects of human biology, e.g. anatomy. In his book B iological Foundations of Language (1967:374ff) he argues that

a. language is innately determined (this supports Noam Chomsky’s proposal of an innate Universal Grammar and is an explanation for the regularity of onset and universal pattern of language development between the second and third year of life)

b. exposure to language is necessary to trigger the development of neurological language processes (he brings up the analogy to the relationship between nourishment and growth: food is the architectural raw material which must be chemically processed before it may enter the synthesis that produces tissues and organs)

c. first language acquisition is only successful if the learner is older than two and has not reached puberty yet (according to him, it is impossible before age two due to maturational factors and after puberty because of the loss of “cerebral plasticity”, i.e. lateralization of the brain already took place) These three postulations are the core of the so-called Critical Period Hypothesis (CPH), a meanwhile common expression in scientific publications which is attributed to Lenneberg although he did not introduce this term himself.

[...]

¹ More accurately, Bornstein declares these five parameters as rather outdated and presents seventeen features which are used recently. However, for the purpose of my term paper the “old” five are illustrating enough and are still the core of the extended register.