ie in native words, except in the word bijzonder ‘special’. It is also written ie in the final syllable of many underived hybrid words (in order to create an adequate spelling for forms with suffixes). In pre-final syllables, one finds usually i in hybrid words. When the class of native words is not distinguished from the class of hybrid words, either the large set of native words will be exceptional, or the large set of hybrid words.
12 Not all diminutive forms introduced in GB 1995 followed this strategy. For instance, the diminutives of deux-chevaux (2CV) and pince-nez (pince-nez) are deux-chevauxtje not *deux-chevootje, and pince-neztje, not *pince-neetje.
13 According to the prize-winning essay by Siegenbeek 1804, one of the founding fathers of present-day Dutch orthography, its relation to pronunciation should be fundamental. Accordingly, Te Winkel’s Phonological Principle has primacy over the Morphological Principle, and the 1954 reform aimed at a phonologically faithful ‘dutchification’ of inflected forms. The result is exemplified by the diminutives in the left-hand column of (3), skister being an exception that was only added in 1995. A principled approach is called for, in which either the paradigmatic uniformity of diner - dinertje etc. prevails, or compliance to pronunciation as in ski – skiester, oma - omaatje, etc. In other words, the hierarchical ordering of the Morphological and the Phonological Principle needs to be re-examined.
14 Cf. GB 2005:94. Observe that FAQ’jes should be FAQ’tjes for those who pronounce the word letter by letter. In Dutch diminutives, the allophony of the diminutive suffix is expressed in the spelling. Such variation will be the inevitable consequence of a spelling based on pronunciation. It is one of the aspects in which an abstract spelling is superior. See Neijt & Schreuder (to appear) on the issue of variation of pronunciation and abstractness in alphabetic writing.
15 However, the stem of some English verbs had been dutchified already in 1995: stressen, ik stres/*stress;scoren, ik scoor/*score (GB 2005: 81) .
16 Since forms such as fonduedeshow that the stem is fondue, the infinitive should have been fondueën in the 1995 spelling. In barbecuen, which has antepenultimate stress, it can be assumed that the e is deleted after an unstressed syllable just as in neuriën.
17 Admittedly, some frozen word groups are written as orthographic words, without spaces between the constituent words. Examples are hogeschool ‘high school’, sterkedrank ‘liquor’ and weliswaar ‘however, admittedly’ (lit. ‘well+is+true’). However, such writing seems to be the exception rather than the rule. Examples of frozen word groups written with spaces are vaste prik ‘business as usual’, hoge hoed ‘high hat’, op grond van ‘based on’, and many, many others.
18 Meester is added to the list of items which were already considered special adjuncts: niet ‘not’, non ‘non’, bijna ‘almost’, oud ‘former’, ex ‘ex’, aspirant ‘trainee, candidate’, adjunct ‘deputy, vice-’, substituut ‘substitute’, chef ‘chief’, kandidaat ‘candidate’, interim ‘interim’, stagiair ‘trainee’, leerling ‘apprentice’, assistent ‘assistant’, and collega ‘colleague’. As we saw in (6e), amateur is no longer part of this list (GB 2005:42).
20 The spelling rules state that optional hyphens may be used in compounds to clarify their meaning, so the spelling rij-instructeur is still possible (GB 2005:36, 53, 78). These optional variants are not included in the dictionary.
21 The following example also illustrates the fact that pronunciation has been a guiding principle of the new hyphenation rules. In the first version of the dictionary, Frankrijk ‘France’ was hyphenated as Fran.krijk because of the pronunciation [frɑŋ–krɛɩk]. However, the word rijk ‘kingdom’ can be recognized in this frozen compound. In such cases the morphological structure used to override the pronunciation, e.g. heel.al ‘universe, whole + all’ with the pronunciation [he–lɑl]. The form Fran.krijk was corrected in the errata.
22 The status of the Donor Principle is unclear. It is less prominent in the unpublished version of the guidelines (the Technische Handleiding) than in the Groene Boekje.
23 GB2005 contains two forms for the two uses of St. Nicholas: Sinterklaas for the mythical bishop, friend and benefactor of all children, who comes from Spain each year and rides the rooftops in early December, doling out gifts and sinterklaas for the many stand-ins who dress up as Sinterklaas, and sit in shops or make house calls on 5 December. There is, however, only one form for Kerstman.
24 Actually, the new rules for acronyms are less straightforward. Capitals are still used (a) when they denote a name: Unicef. Acronyms of fewer than four letters and acronyms of four letters which denote a public institution, union or political party are written with capitals only (but cf. Sdu for ‘Staatsdrukkerij en –uitgeverij’, PvdA ‘Partij van de Arbeid’). (b) Capitals are used when they occur in the donor language, and the acronym is not considered common. Examples are English RAM, ADSL, or German GmbH (the capitals in the acronym remain althoughwe no longer write German nouns with capitals). (c) Capitals are used when the acronym denotes a law, resolution or government scheme VUT, WAO (exceptions are possible under the Donor Principle: AMvB, Wajong). (c) And finally, capitals are used when they denote an illness: BSE, ME(unless the acronym is considered common: aids).
25 In the editions 1995 and 2005 of the Groene Boekje, the spelling principles are less prominent than before. Earlier spelling dictionaries referred to the Grondbeginselen, ‘Fundamentals’, a book of more than 200 pages (Te Winkel 1865). The spelling dictionary of 2005 deals with the principles in only two pages.
26 Meester and oud are only special adjuncts when they mean ‘master’ and ‘former’, respectively. In meesterbrein “the meaning of meester is weakened” (GB 2005:42), and oudgast does not mean ‘former guest’, but has an idiosyncratic meaning ‘someone who lived in Indonesia’.
27 Even when the so-called inconsistency is due to a rule that can be used consistently, as is the case in havoër and vwo’er. Such examples show that one expects that two closely related words are spelled the same way.
28 But notice that judgements vary. One may use Cockney etc. non derogatorily, as well as one may use the form Aussie, dubbed derogatory by the dictionary, as a jocular nickname for people from Australia. The form Jappen may be used simply as a shortened, colloquial form of Japanners ‘Japanese’ in Dutch.
29 Core and periphery need to be distinguished empirically. Presumably, the choice of letters and the couplings between letters and sounds belong to the core, and all the rest (capitals, abbreviations) is periphery. Relevant data can be found in psycholinguistic experiments like those described in Schreuder et al. 1998, Bosman & Van Hell 2002, Frisson & Sandra 2002, Van Heuven 2002, etc. Such experiments could shed light upon the effects of variation in spelling on the efficiency of reading as well.
30 Observe that we evaluate the complexity of the spelling system in terms of numbers of rules and numbers of basic distinctions within these rules. Our conclusion differs from Verhoeven (2007:34), who claims that the new spelling will be easier to learn in certain respects, because there are fewer exceptions. We agree that a set of rules without exceptions will be easier to learn than the same set of rules plus a number of exceptions. However, the spelling of 2005 added rules but did not succeed in eliminating the exceptions. We think that a system of coherent general guidelines which allow for variation in areas that are difficult to express in alphabetic writing, will be easier to learn and use than a less coherent system with many detailed rules. In other words, we think that where no coherent treatment is available, some room for variation is to be preferred over strict regulation. Of course, empirical evidence is needed.
31 Readability is mentioned, however, in the unpublished Technische Handleiding to motivate (a) dutchification of hij skiet (from: ski); (b) the use of dots in some abbreviations; and (c) the optional insertion of hyphens between the words of a compound.
32 The advantage of writing compounds as words, without spaces, is that it expresses scope distinctions. Pronunciation makes you wonder about the interpretation of you’ve got only ten in Bob Dylan’s text about the man in the trenchcoat who wants eleven dollar bills. In Dutch orthography, the difference between vuilegrondaffaire ‘affair about dirty soil’ and vuile grondaffaire ‘nasty affair about land’ can be expressed in writing. Observe that this strategy may also prevent garden paths. The multiply ambiguous sentences time flies like an arrow and fruit flies like a banana would lose some of their many interpretations if the first two words, if and only if interpreted as a compound, were written as timeflies, as can be done with fruitflies. The English way of spelling compounds is notoriously difficult for technological language applications, cf. Isabelle & Bourbeau (1985:21). The Dutch way of spelling compounds is superior in this respect.
33 Verhoeven (2007:38) concludes that pros and cons are balanced. He claims that the differences are small, and that the new spelling is simpler for the writer and more difficult for the reader. In our opinion, the spelling of 2005 illustrates better than the spelling of 1995 a new approach towards the relation between spoken and written forms. The spelling law of 2005 marks the demise of spelling along the lines set out by De Vries and Te Winkel, not just because of the cancellation of the old law that referred to their book on basic principles. More importantly, the new spelling no longer aims at formulating general principles, from which the detailed rules are derived and which offer a solid base for choosing among variants. Instead, the slogan “a problem with a solution is no longer a problem” (Een probleem met een oplossing is geen probleem meer, title of an interview with the secretary of the Taalunie, Linde van den Bosch, de Standaard October 17, 2005) illustrates the new approach: detailed rules, no principled method to handle fundamental issues, different types of consistency within small sets of words.
34 For laymen, amateurs, the spelling should be a handy toolbox, not a millstone, a loodgieterstas ‘plumber’s toolkit’ (Van de Laar 2006).