David Leonardi; 2018 (self-published); pp 194
This review article first appeared in: Mark’s Bookshelf, Skeptical Intelligencer, Winter 2018
Reproduced with permission
Back in 2014, I reviewed in this forum (in two pieces) David Leonardi’s last book, published in the previous year: Egyptian Hieroglyphic Decipherment Revealed: a Revisionist Model of Egyptian Decipherment Showing Evidence that the Ancient Egyptian Language and the Ancient Hebrew Language are Closely Related. (For more on Leonardi’s older work, see the relevant sections of my 2013 book Strange Linguistics.) The main focuses of Leonardi’s thought are: a) the allegedly close relationship between Egyptian and Hebrew (as in his 2013 book); b) (associated with a)) the alleged misinterpretation of Ancient Hebrew by medieval and later scholars and of the Egyptian language by the C19 decipherers; c) the allegedly close relationship of early Hebrew with a recent ‘Proto-World’ (the supposed single ancestor of all human languages), as might be suggested by a literal reading of Genesis; (d) the alleged methodological (especially statistical) errors of most mainstream historical linguists in arriving at their contrary view that Proto-World (or multiple ancestor languages) long pre-dated all known languages, was therefore in no way especially close to Hebrew or any other known language, and cannot in fact be reconstructed; Leonardi for his part holds that in contrast the figures support the view that such deep-time relationships can be demonstrated, at least to a degree, and that his specific findings favour his theory that early Hebrew was indeed close to Proto-World. (Compare the ideas of Isaac Mozeson, discussed in my ‘Language On The Fringe’ column in Skeptical Intelligencer 21:2 (2018) and referred to here by Leonardi on pp 86-87.)
In this last respect, Leonardi aligns himself (in my view, without an adequate grasp of the issues) with a minority group of C20-21 professional historical linguists who hold that relatively recent dates for Proto-World should be accepted and thus that Proto-World can (perhaps) be reconstructed in part, using methods rather like those employed by Leonardi (albeit more sophisticated; for example, they rightly focus upon the oldest available attested or reliably reconstructed forms, which Leonardi himself sometimes omits to do; see for example his discussion of English and Thai on p 61 of this book). (It is agreed by all parties that the more reliable ‘comparative’ method of establishing historical relationships between languages and language ‘families’, which crucially involves systematic correspondences between phonemes and forms, cannot itself be extended far enough into the past to be used here. Leonardi refers to this contrast of methods on pp 20-21, 32-33.) The best-known figure in the minority group is Merritt Ruhlen, and Leonardi refers to his work and that of his followers (for example on pp 29-32, 45-47), and also (especially on pp 55-66) to that of Alexis Manaster Ramer, William H. Baxter, Lionel Bender, etc.; some of these linguists are in fact theoretically closer to Ruhlen than are others. (Like some other non-mainstream writers, Leonardi at one stage appeared to be regarding Ruhlen as core-mainstream and reliably authoritative; but, while he still endorses Ruhlen’s stance against core-mainstream opposition, he now seems more aware of Ruhlen’s perceived ‘maverick’ status.) On pp 12-13, 20-21, 32-33, 46-47, 66, etc., Leonardi critiques the mainstream position referred to under (d) above, makes some very forthright and inadequately defended statements (for example, on p 12 he rejects as too high the ‘rule-of-thumb’ figure of 5% normally given for chance linguistic similarities, without here presenting or foreshadowing any argument), and accuses the leading linguists in question of fatally misunderstanding or misrepresenting the ideas of the minority group.
Leonardi’s titles for his works, and indeed his self-appraisal generally, are typically dogmatic, not to say bombastic, and the title and tone of this his latest book is in the same vein. It is important to remember in this context that while Leonardi is quite widely-read in linguistics he has no formal background in the discipline and frequently displays confusion – and is unwilling to be corrected on his errors. He disputes or ignores clearly valid points made in correspondence, for instance on the morphology of Greek (as he admits on p 87 of this book, he is still insufficiently familiar with this language) and on the thought of mainstream linguists and in particular that of the leading morphologist Peter Matthews (which he misunderstands); and he does not refer to reviews even by way of attempted rebuttal (because most mainstream linguists would not be aware of him, he presumably receives feedback mainly from his own allies, on whom see below). Furthermore, his exposition of his own ideas has often been obscure, and this problem arises again in this present work (as is agreed by another reader whose help I enlisted; see below). Leonardi is in fact merely the founder and leader of an idiosyncratic online discussion group with the misleadingly mainstream-sounding name ‘Historical Linguistics’. His self-confidence and his status as a ‘big fish in a small pond’ should not deceive unwary readers, especially those who have ideological reasons for hoping that he is correct about Hebrew.
My main focus here is upon the Preface (pp 5-8) and Chapters 1-3 (pp 9-78), where Leonardi presents most of the novel material in this book.
In the course of his argument in these chapters, Leonardi (to his credit) attempts more overt use of statistics than do most other such authors (see especially pp 34-36, 47-65, 75-78). But, in considering the statistical likelihood of word forms in apparently unrelated languages being themselves related (either by shared ‘genetic’ origin or through contact and degrees of bilingualism), he seems to be trying to assess probabilities of accidental similarity in an egregiously ‘coarse-grained’ manner, ignoring the point (regarded by linguists as crucial) that the precise likelihood of accidental (unsystematic) similarity between pairs of words, as opposed to genuine connectedness of one kind or another, will depend upon a number of factors (this is why the ‘rule-of-thumb’ referred to above can be no more than that, contrary to Leonardi’s perception). These factors include:
(a) the degrees of (unsystematic) phonological and semantic similarity between words which are required if they are to be counted as shared (there is obviously a degree of subjectivity in many such cases, but writers such as Leonardi have a bias in favour of ‘finding’ such similarities where they bolster their own theories and not recognising them where they would appear to support alternative non-mainstream theories; this pattern of interpretation is predictably common in non-mainstream work produced by authors with ideological or personal ‘axes to grind’ – many of whom are simply unfamiliar with the requirement of systematicity)
(b) the phonological and semantic systems of the relevant languages (for example, how far the major groups of phonemes are shared or have close systemic equivalents)
(c) the lengths of the words (for example, if two languages not known to be connected share a very short word-form such as [sa] with the same meaning, this could very well be accidental, whereas if they share the form [tolpesveblig], again with the same meaning, or with transparently related meanings, this is less likely); in fact, in cases involving longer words, Leonardi tends to focus on sub-components of the words, typically the initial few phonemes and/or the consonants rather than the vowels (this latter, like some of Leonardi’s errors regarding Greek, is to be expected of a non-mainstream Hebraicist with a limited knowledge of non-Semitic languages); see for instance his example on p 50 involving roots featuring the consonants [r] and [d]; he thus obscures this specific factor
(d) the cross-linguistic frequency of the sounds and sound-sequences in question (very widely-shared sounds such as [e], [s], etc. or common sound-sequences such as [til] or [po] are more likely to be shared by chance than sounds and sequences found in relatively few languages)
(e) vocabulary size in the various languages in question (which Leonardi admittedly does discuss, albeit only very briefly and with a focus on samples of vocabulary, on p 60)
The (anonymous) second reader whose help I enlisted, who is a professional statistician (but, it should be noted, not specifically trained in linguistics), agrees with these observations. He also notes that Leonardi’s use of statistics, more generally, appears naïve. Judging from, for example, his laboured explanation of the Poisson distribution on p 49 (where he also appears not to define ‘n’), Leonardi seems, in fact, to be making some unjustifiable assumptions and statistical errors. His critiques of mainstream historical linguistic methods in this area, and his own statistically-based proposals, therefore appear at best very doubtful.
Leonardi might hope to develop a stronger case for his claims if he placed more emphasis on grammar (morphology and syntax). But (as adumbrated above) grammar (especially that of languages other than Hebrew) is not his strong point, and in fact (like most non-mainstream historical linguists) he has little to say about it, focusing mainly on lexical phonology (words and their pronunciations), where his problems as set out above arise.
On p 48, while ‘building up’ to some statistical comments, Leonardi introduces a four-way typology (similar to typologies presented in his earlier work) involving various kinds of (accepted or purported) correspondence between word forms which are regarded as (‘genetically’) unrelated or are liable to be so regarded by those coming to the material. Confusingly, he includes here (as ‘Category 3’) ‘actual related correspondences which do not count as a correspondence because of phonetic change having obscured the resemblance’; that is to say, it would probably not have occurred to analysts of any persuasion to treat them as possibly related if it were not known that they were, and Leonardi reasonably excludes them from his further analysis. But such forms – most of them in fact ‘genetically’ related – are very typically known to be related only through analysis of systematic but opaque phoneme-correspondences in relatively shallow time (see above), and they are essentially irrelevant to Leonardi’s discussions of putative deeper-time correspondences. (One such example is the surprising historical identity of English fig– in figment and –dise in paradise, both coming from a common Indo-European root meaning ‘build’.) In any event, Leonardi’s subsequent discussions of specific alleged correspondences (for example on p 50) are often obscure and/or confused.
In Chapters 4-9 (pp 79-172), Leonardi takes up again (and develops somewhat) i) his discussions of Hebrew and Egyptian and their supposed lexical links with English and a wide range of other languages, and ii) his confused/confusing theories of derivational morphology, vocabulary ‘strata’ (‘basic’ versus ‘advanced’/‘abstract’, different languages of origin; he introduces these notions here on p 13) and ‘borrowing’ – much of which forms the main content of his earlier books. Because of his failure to engage with criticism of these earlier books or to alter his approach to these matters in any significant way, I do not propose to comment here on these chapters.
In Chapter 10 (pp 173-184) Leonardi discusses the supposed role of ‘Koine Greek’ (as in the New Testament) in the development of Hebrew phonology in the period in question (from 300 BCE to Late Antiquity); but his confused ideas about Greek and his extreme and unjustified view that the sounds of all ancient languages represent variants on the phonemes spelled with the 22 letters of the Hebrew abjad render much of this discussion incoherent.
Leonardi sums up in a brief Conclusion (pp 185-188) which predictably (and not unreasonably) adds little to his arguments. Once again, it cannot be said that overall he has made out a convincing case for his ideas.
There is occasional loose referencing (for example to an unattributed Youtube video, on p 36). And not all of the scholarly works cited in the body of the text appear in the Bibliography. These faults should be remedied in any new edition. But the book would need to be totally re-worked if it were to have any real significance.