arroba Email

Robert Pennock’s Book of Babel

A Review of Robert Pennock's Tower of Babel (Part II)

(continued from Part I)

If the observation of growth and decay in nature has fostered both the classical model of human history as decline and its modern opposite, the myth of progress, we should not be surprised that both have strongly influenced our view of language development. In fact, it is only in relatively recent years that linguists themselves have succeeded in freeing themselves from such mindsets.

Looking back on Schleicher’s notion of Sprachenverfall, we might note his striking comments about the native language of Robert T. Pennock:

“The [English] language may have preserved the Anglo-Saxon type but is one of the most eroded and, in grammatical endings, poorest languages of our [Germanic] branch. Most of the originally Germanic words have even degenerated into monosyllabic forms—at least in pronunciation, which here is alone decisive, and this has deviated considerably from the written form, thereby providing clear proof of how quickly the language of an historically and literarily significant people can go into decline.” (ibid., p. 231)

In 1894, the renowned linguist and English language scholar Otto Jespersen published Progress in Language, in which he seeks to refute Schleicher’s views, arguing that the radical loss of inflectional distinctions, as seen in the history of English, indicates not degeneration but rather improvement:

“The direction of movement is towards flexionless languages (such as Chinese, or to a certain extent Modern English) with freely combinable elements; the starting-point was flexional languages (such as Latin or Greek); at a still earlier stage we must suppose a language in which a verbal form might indicate not only six things like cantavisset [i.e. third person, singular, active, pluperfect, subjunctive, Latin sing CD] but a still larger number, in which verbs were perhaps modified according to the gender (or sex) of the subject, as they are in Semitic languages, or according to the object, as they are in some American Indian languages. It is indeed highly characteristic of the primitive mind, and a subject of constant astonishment to those who study the languages of savage races, that a thing by itself cannot be conceived or spoken of: it is an utter impossibility for a savage to think of ‘knife’, for instance, by itself; his power of abstraction is not sufficiently developed; but he can perfectly well say, ‘give me that knife’ or ‘he plunged the knife into the hart’.”8

Jespersen was, unlike Schleicher, a consistent and ardent Darwinian. In his introduction to the reprinted volume of Progress in Language, the late James D. McCawley, former president of the Linguistic Society of America, deplores Jespersen’s “seemingly racist descriptions” but, with a gentleness quite atypical of these politically correct times, attempts to put the lapse in historical perspective. McCawley’s description of Jespersen’s Darwinism is likewise less than critical, though in one passage we find a key to the entire issue:

“There are several ways in which one can dispute Schleicher’s view of his flexional type as the pinnacle of linguistic evolution. In the 20th century, the most common response to that view has been to dispute that there is any pinnacle to linguistic evolution and to hold that there is no particular direction to linguistic evolution [emphasis added], with languages of any of Schleicher’s types being able to develop in the direction of either of the other types.”(ibid., p. xi)

According to Darwinian biology, random mutations combine with natural selection to produce evolutionary change: the one is environmentally unconditioned, the other conditioned. Now the first part of this analogy clearly applies to language, in that sporadic and even contradictory changes occur that are not explainable by environmental factors. The English word ask, for example, which goes back to Proto-Germanic *aiskojanan, underwent metathesis (the reversal of segments) in Old English to become acsian. S and c were soon reversed again, so that in standard English today we say ask. The rival form survived in various dialects, however, and, to judge from varieties of African-American speech, seems to spring up again spontaneously: aks.

The problem here is that there is nothing in language change comparable to “natural selection” to determine the victorious form. Observation of known linguistic change, cross-linguistic data, and the principles of articulatory phonetics allow us to make tentative generalizations about the internal dynamics of linguistic structure, but even so the cause-effect relationship is often tenuous.

While there are universal “tendencies,” the details vary significantly from language to language. In regard to ask vs. aks, a phonologist may rightly say that non-initial ks is somewhat less “marked” than non-initial sk, while initial ks is far less common than initial sk. Yet Greek, for example, has preserved Indo-European *qs-, as ks in xýlos “wood” and xénos “stranger”. (In English borrowings, we find it pronounced [z], as in xylophone and xenophobia, but in German and French, they are treated as learned words and pronounced as in the original.) In Sanskrit, we even find an example of kš resulting from a metathesized *sk: kšar- < *skar- “flow”.

Languages are, of course, altered through contact with each other, and these may be regarded as the external dynamics of language change, but even so, such “evolution” (to use Pennock’s conveniently slippery term) is entirely limited to their lexical and semantic components. The Norman Conquest accounts (in part) for the replacement of (Old English) onfon by receive, and the advancement of science helps to explain the modern meaning of ether; on the other hand, such changes in English as the Great Vowel Shift or the reclassification of some strong verbs as weak verbs (e.g. helpan, healp, holpen > help, helped, helped) cannot be attributed to external causes.

Linguists in the past have often failed to distinguish strictly linguistic from the sociolinguistic (historical, cultural, and political) factors that may determine the overall fate of languages. Irish Gaelic, for example, has long been yielding to the onslaught of English, but when modern scholars discuss specific factors, they mention British imperialism or the impact of the Great Famine. None would seriously suggest that Gaelic speakers were handicapped by observing verb-initial word order,9 morphologically marked case distinctions, and grammatically conditioned spirantization and voicing of word-initial consonants.

Similarly, contemporary scholars, particularly those in the United States, would universally (and indignantly) reject the following claim made in 1865: “As we can now perceive, certain peoples, such as the North American Indian tribes, are unfitted for historical life because of their endlessly complicated languages, bristling with overabundant forms; they can only undergo retrogression, even extinction.” The source is Über die Bedeutung der Sprache für die Naturgeschichte des Menschen [On the Significance of Language for the Natural History of Man] by August Schleicher.10 Had Schleicher lived a bit longer, he might have seen that the fate of the Algonquian languages (e.g. Cheyenne and Arapaho) had nothing to do with their polysynthetic morphological structure and everything to do with land, gold, guns, and smallpox.

Though language is, to varying degrees, linked to ethnic and cultural identity, its ties to gene pool are far more tenuous, for it is ultimately an abstract, non-material entity. At the same time, it is not a purely cultural artifact, being only marginally manipulable by the conscious human will. To employ Schleicher’s terminology, it partakes of both Natur and Geist, without belonging entirely to either.

Not surprisingly then, the history of individual languages does not conveniently coincide with that of their speakers. Conquerors may impose their language on the conquered, but there are not a few examples of the reverse: the Scandinavian-speaking Normans, who came to speak French: the Manchus, who abandoned their language for Chinese; the Israelites who, some scholars believe, spoke Aramaic before adopting Hebrew, the language of the Canaanites. There are even instances in which a language thrives even when its original speakers disappear or are absorbed. Aramaic, the lingua franca of much of the ancient Middle East, is a striking case in point.

Even the death or extinction of languages cannot be understood except in a loose metaphorical sense. Latin, the “dead language” par excellence, though no one’s first language, is very much a living means of communication, with many more competent users than the speakers of hundreds of so-called extant languages. The preservation of Sanskrit and Hebrew as sacred languages and their subsequent revival as spoken languages again point to the distinction between linguistic and biological entities. To drive the argument to an admitted extreme, languages “live on” in some sense, even when all that is left is a handful of borrowings.11

The ultimate irony of Pennock’s attempt to bolster Darwinian theory by pointing to linguistic “evolution” as an analogous process is that the phenomenon of language, both in its similarities to living organisms and in its differences, ought to be a source of perplexity for those in his camp. Pace Pennock, the direction that linguistic science has taken since the 19th century has generally been not towards the Darwinian model but rather either away from it—or, at the least, towards a position of embarrassed equivocation.

Perhaps referring to well-publicized experiments intended to test the language-acquisition capabilities of non-human primates, Pennock confidently assures us: “Though the study of animal language remains controversial, it is fair to say that recent evidence has further supported Darwin’s argument that the differences are of degree rather than of kind.” (p. 126) In fact, linguists, particularly of the Chomskyan school, have long been skeptical, if not scornful, regarding such research, and, if anything, “recent evidence” (despite all the media hype) has pointed in the opposite direction.

TOB’s numerous linguistic howlers strongly suggest that its author failed to take advantage of the resources at hand in the renowned linguistics department of the University of Texas at Austin. Having written with wide-eyed wonder of his discovery that there is variation within languages, for example, that Texans say “howdy” and “y’all,” he writes:

“The plural ‘you’ also used to be distinguished from these familiar singular forms [thou and thee] as an honorific form to be used when speaking to a social superior, but linguistic historians note that this began to change in the seventeenth century when Quakers rejected the honorific use of the term as a ‘denial of the equality of all men’…” (p. 128)

Pennock displays a touching faith in the power of his Quaker ancestors to alter the English language, but the truth of the matter is that you had been the unmarked form of address for at least half a century before George Fox and his followers began to “thou” everyone. Even in their day the usage was archaic or dialectal.

Again breathless with excitement, Pennock informs us that Old English translations of the Bible are significantly different from their modern counterparts, which he further compares with the Latin version. Focusing on the word for “father” in the Lord’s Prayer, he comments:

“Kinship terms are among the words most highly resistant to major change, as we see above in the variations in the word ‘father’ The transition from the Latin reflects a regular transformation of the ‘p’ to ‘f’, and the ‘t’ sound to the softer ‘d’ and then ‘th’.”(p. 131)

My professor for Indo-European and Germanic would surely have winced at this description and wrapped its author on the knuckles. Most glaringly, he repeats the erroneous notion, apparently widespread among the linguistically naïve, that English is directly descended from Latin. The shift from Indo-European *p to Germanic *f is estimated to predate the Latin translation of the Bible by nearly a millennium. *t did not directly change to *d; because of the original non-initial accent in Germanic, it became (voiced) th, then in West Germanic d. Moreover, there is nothing “regular” about the further “transformation” of d to (voiced) th: it occurs only sporadically in words in which there is a following r, cf. weather vs. fodder.

Again unwittingly scratching his nails on the linguistic blackboard, Pennock refers to lexical borrowing or blending as
“hybridization.” 12 He admits that it occurs on such a “massive scale” that it points to “significant differences between how languages and biological entities can evolve.” (p. 139) Ironically, linguists generally deny the existence of genuine hybrids, which they define as languages whose structure as well as vocabulary is heterogeneous. Thus, Vietnamese, for example, whose vocabulary consists mostly of Chinese loanwords, is nonetheless classified as Austro-Asiatic, not Sino-Tibetan. Recently, thanks in part to the study of pidgins and creoles, the possibility of structural blending has come to be considered more favorably, notably in regard to hypotheses regarding the origin of the Japanese language. Yet even so, hybridization in the true sense is, at the very least, extremely rare.

Pennock clearly fails to realize just how severely he is undermining his own argument here. In the biological world, hybrids typically result in sterile or non-viable offspring. In the case of languages, whose ultimate survival, as explained above, has nothing to do with their internal structure, such is irrelevant. Moreover, languages, again unlike plants and animals, have no “life imperative” of their own. To imply that linguistic change is comparable to the natural “adaptation” to which living creatures are subject would be as absurd as to suggest that the hems of skirts rise and fall as a part of an instinctive defense mechanism against trousers.

Another example of Pennock’s inadequate understanding of linguistic principles is seen in his comparison of natural and “designed” languages such as Esperanto:

“Despite their derivative nature, these designed languages are notably different from naturally occurring ones. Immediately apparent is that the syntactical rules of artificial languages are completely regular—designed that way for ease of use—in stark contrast with the irregular verb forms, inconsistent rules for forming plurals, haphazard assignment of nominal gender, and so on that are the bane of every student of natural languages…Natural languages…,despite the commonalities of form and the Chomskyan expectation of an underlying universal grammar, are full of unexpected twists and bizarre exceptions to every rule…We expect that languages that are designed and specially created would be structured in a simple and regular manner, as is Esperanto, rather than be the jerry-built jumble that is natural language.” (p. 146)

Reading this, I was reminded of Flambeau in Chesterton’s “The Blue Cross,” who, disguised as a priest, remarks to Father Brown: “Well, I think that other worlds may perhaps rise higher than our reason. The mystery of heaven is unfathomable, and I for one can only bow my head.” The detective priest is not taken in and explains how he has seen through him, adding: “But, as a matter of fact, another part of my part, too, made me sure you weren’t a priest…You attacked reason…It’s bad theology.”

Pennock seems to be attempting to bow his head to the mystery of language by emphasizing its unpredictability, but in so doing he is merely engaging in “bad linguistics.” Even if we were to play along with his strawman of “specially created” languages, his “expectation” about simplicity and regularity is merely his own philosophical conjecture. Where he sees a “jerry-built jumble,” the linguist finds an underlying (and demonstrable) order. Moreover, complexities that may indeed be “the bane” of beginning second-language learners are not so for native speakers.

Most importantly, while the concept of “universal grammar” is not without ambiguities and controversies, Pennock’s misunderstanding of it is clear. The argument made by the universal grammarians, including Chomsky, is that human language is not, for all its diversity, open-ended but is rather conditioned and constrained.13 Steven Pinker, whom Pennock approvingly cites as an evolutionist, is also a leading universalist. Again, we see an argument unwittingly undermined by its own author.

What Pennock calls “syntactical rules” are, in fact, morphological rules, which in Indo-European languages are complex and which language designers such as L.L. Zamenhof, who first constructed Esperanto, have sought to simplify. Morphology is, however, only part of what makes individual languages distinct and thus (initially) inaccessible to the non-native learner. Other grammatical and semantic categories may make them seem just as arbitrary. Proponents of Esperanto can rightly point to its morphological transparency, but, particularly if they are speakers of Western Indo-European languages, may fail to realize that for speakers of non-European languages, such problems as the use of the definite article and singular-plural marking involve semantic distinctions that are far from obvious.14

In attempting come up with the linguistic equivalent of genetic inheritance, Pennock argues that children acquire their mother tongue through imitation or copying and that this is somehow analogous. He continues:

“Think first of mutation. The monks who spent their days prayerfully writing out the Bible had tight quality control, but if one happened to make a minor error, that mistake could be carried on when the next illiterate monk made his own copy from the original, and in this way a ‘mutation’ would be inherited by subsequent generations in that lineage of copies.” (p. 138)

Here, to justify one analogy, Pennock reaches for another. The result is more bad linguistics—and bad philology.

First, scribal errors that result in a non-existent or, in the context, nonsensical word are easily corrected. Thus, for example, “The boy hissed the girl” will eventually be “caught” and rewritten as “The boy kissed the girl.” Mistranscriptions that are much more likely to endure are those written not by the mindlessly illiterate but rather by those who are so aware of what they are transcribing that they unconsciously adapt it to their own phonological and grammatical speech norms. An English speaker in whose dialect ferry and fairy are distinguished is less likely to write “I took the fairy across the bay to San Francisco” than is somehow who pronounces the two words in the same way; a trained Iranian typist with only a minimal command of English is even less likely to make the mistake. Similarly, if we find helped for holpen in a medieval manuscript, we assume not that a sleepy monk happened to write “e” for “o” and “d” for “n” but rather that in the English that he spoke, a grammatical change had taken place.

The suggestion that children learn their mother tongue through passive “copying” flies in the face of both Chomskyan theory and all mainstream assumptions about language acquisition. If the “illiterate monk” analogy were to hold, then we would expect toddlers to make mistakes comparable to confusing an “h” for “k.” In fact, what makes their errors so interesting is that while they cannot be specifically predicted, they can typically be explained in terms of general linguistic patterns.

Again, the phenomenon of metathesis provides us with examples that are particularly appropriate, in that they illustrate both vacillating (non-directional) change and phonetic principles. In the history of English, we find numerous cases of “r” shifting positions with an adjacent vowel. Proto-Germanic hors “horse” is found just as is in Modern English, but in Old High German it became hros (> Modern German Roš). Old English, brid “small bird” and thrid- have become bird and third; on the other hand Old English thurh and beorht have become through and bright, cf. the dialectal variants purty for pretty and prevert for pervert. (The former, the cause of much mirth among the more genteel viewers of the film Dr. Strangelove, undoubtedly smacks of unlettered bigotry, but from a longer, linguistic perspective, we see that the moveable “r” in the prefix goes back to Proto-Indo-European.)

While variation can lead to a decisive change in form,15 the crucial difference between biological and linguistic “mutation” is again that innovations in language are not subjected to the test of environmental adaptability. One can, of course, claim that “pervert” wins out over “prevert” because of the strong social urge not to appear to be a bumpkin, but that argument clearly puts the cart before the horse.

Phonetic factors, notably ease of articulation, are often cited as the single most powerful motivating force in sound change. Yet we are still left to account for the great diversity of phonological systems—and of their history—among the languages of the world.

English-speaking children sporadically pronounce animal as [aminal]; Japanese-speaking children, as I can attest from my own parental experience, are known to render kodomo “child” as [komodo]. The phonetician can account for both cases by saying that pronouncing the bilabial before the dental, i.e. “m” before “n,” “m” before “d,” is marginally easier than the reverse. The more interesting question, however, is why the “progressive” speech of young speakers is generally so ephemeral. One piece of evidence pointing to the answer is the looks of suspicion adults typically receive when they deliberately imitate such errors. Language-learning children, far from being like Pennock’s passive monks, are, as were, well aware of the “manuscript” with which they are dealing and are struggling to master it.

Finally, I refer to what follows the passage cited above in order to point out a blunder, which, though hardly central to Pennock’s overall argument, is revealing of his book’s overall tone:

“Such a mistranscription or mistranslation might have been what led to an amusing passage in the Bible saying that it is easier for a camel to pass through the eye of a needle than for a rich man to enter the kingdom of heaven. Although the surreal image of trying to thread a camel through a needle does quite dramatize the difficulty, I have always found the proverb a little odd. The metaphor made more sense after I heard that the term probably meant to be ‘rope’, which differed from ‘camel’ in the original language by just one letter.” (p. 139)

The “proverb” here is the well-known admonition of Jesus, which appears not in one passage of the New Testament but rather in three of the four gospels: Matthew, Mark, and Luke. Jesus also refers to camel-swallowing Pharisees. The theory that the word originally intended was “rope” is likewise familiar to the more Biblically literate, but as John L. McKenzie, S.J., comments:

“The violence of both figures led a number of critics to suspect that the original reading was Gk kamilos, cable (pronounced like kamelos, ‘camel’, in later Gk). The MS evidence, however, supports kamelos fully; and the violence of the metaphor is less in ancient Near Eastern speech than it would be in modern speech. Hyperbole is common in oriental languages.”16

The unwitting perpetuation of a flawed or outdated conjecture is a risk that all of us run. Here, however, as in so much of the rest of his book, Pennock is attempting to display a broad range of knowledge that he obviously does not possess. Perhaps lulled into a false sense of confidence in dealing with a quaint work of folklore, namely the Bible, and one of its more “amusing” passages, he displays an arrogance that so often typifies today’s self-appointed apostles to the benighted.

1 After Babel: Aspects of Language and Translation, p. 56. Oxford University Press, New York and London

2 Science of Thought, vol. I, p. xi. (London: Longmans, etc.; New York: Scribner’s Sons, (1887).

3 Die Sprachen Europas in systematischer Übersicht, p. 2, Amsterdam Classics in Linguistics, Amsterdam/Philadelphia: John Benjamins Publishing Company, (1983) Translations here and below, unless otherwise indicated, are my own.

4 Among politically correct linguists, the principle of laissez-parler does not exclude those forms of prescriptivism consistent with the egalitarian agenda. The teaching of Standard American English to inner-city African-American children, for example, is much more likely to be condemned as linguistic fascism than the insistence that aged males abandon the Miss/Mrs. distinction for Ms.

5 An exception must be made for those who devote their linguistic skills to the theory and practice of Biblical translation and, by extension, to evangelism, activities frequently looked upon by their secular brethren with suspicion and even overt hostility.

6 Note that to form the Latin sentence, one must know the correct nominative plural form of piscis, a third declension noun, the correct ablative singular form of fluvius, a second declension noun, and the correct third person plural form nare, a first conjugation verb. In the Korean example, the various grammatical markers, though not strictly invariable, are transparent and predictable. There are no inflectional classes.

7 Typology, having long languished under a cloud, in no small measure because of the contemporary emphasis on universal grammatical principles, has regained respectability within the broader framework of contrastive linguistics.

8 Progress in Language: With Special Reference to English, p. 348, Amsterdam Classics in Linguistics, Amsterdam/Philadelphia: John Benjamins Publishing Company, (1993)

9 More precisely, the order is verb-subject-object: Díolfaidh m’athair an tigh mòr “My father will sell the big house” (lit. Sell-will my father the house big). In English, by way of contrast, we find subject-verb-object (SVO) and, in Japanese, subject-verb-object (SOV). Only about 10% of the world’s languages show a consistent verb-initial word order. The tendency in some VSO and VOS languages to topicalize the subject and move it to the front has led some linguists to claim that there is a general linguistic tendency towards subject-initial order. This could well be true, but it is by no means implied that this has anything to do with the overall decline of Celtic languages such as Gaelic.

10 in Linguistics and Evolutionary Theory: Three Essays by August Schleicher, Ernst Haeckel, and Wilhelm Bleek, edited by Konrad Koerner, Amsterdam Classics in Linguistics, Amsterdam/Philadelphia: John Benjamins Publishing Company, (1983). The translation here is by J. Peter Maher.

11 The words maize and potato, for example, derive ultimately from Taino, a Carribean language whose speakers were annihilated with the arrival of Columbus.

12 He offers as one his “favorite examples” Anglo-Japanese—Makku-do-naru, which means to eat at McDonalds. The only such term I or my native teenage sources have heard is “makkuru,” which, like most such trendy slang proved to be highly ephemeral and is no longer in common use.

13 Languages vary, for example, in how they form questions, but none lacks a device for distinguishing them from declarative sentences. “The children were playing with matches”/”What were the children playing with?” The variation of word order that we see in English is not universal, nor is the considerable flexibility of English syntax, which permits the questioned element to be deeply embedded in the sentence: “What did Mrs. O’Reilly insist to the headmaster that she saw the children playing with.” Yet even in English, the sentence “Mrs. O’Reilly scolded the children who were playing with matches” has no interrogative counterpart that focuses on “matches” in the relative clause as the questioned element: *”What did Mrs. O’Reilly scold the children who were playing with?” If, as the universalists argue, such constraints apply cross-linguistically, they tell us something about the true grammar of human language as a whole. (Here and in the following note, an asterisk marks ungrammaticality.)

14 Japanese, for example, is one of many languages that do not grammatically distinguish countable and uncountable nouns, which leads Japanese speakers of English to make such non-morphological errors as *”I haven’t the needed informations to answer your question about the equipments in this plant.” Furthermore, languages in which the countable-uncountable distinction is made may vary as to which is which. A French speaker of English, for example, may make the same mistakes in the sentence given.

15 The motivation for thurh > through and beorht > bright may have been the elimination of the consonant cluster -rh. Any possibility of complete reversal was eliminated with the loss of h.

16 Dictionary of the Bible, p. 116. Collier Books, MacMillan Publishing Company, New York (1966)