Word-signs and sign groups in the Kohau Rongorongo script of Easter Island Russian State University for the Humanities, Institute for Oriental and Classical Studies (Moscow) Abstract: There is an inclination in rongorongo studies to consider the Easter Island script as an essentially syllabic writing system. This assumption is problematic from a typological point of view. The purpose of the paper is to show that word-signs are present in the script referring to combinatorial properties of signs only. For this I propose methods that permit to identify particular signs as word-signs. The use of some signs in isolation, their dependence on context and phonetic complements suggest that kohau rongorongo is a heavily logographic script with a solid phonetic component, apparently lacking semantic determinatives. Bio: Albert Davletshin was born in Norilsk, Russia in 1976. He completed his Ph.D. thesis "Palaeography of Maya hieroglyphic writing" at Knorozov Centre for Mesoamerican Studies, Russian State University for the Humanities, Moscow (2003). He works on Eastern Polynesian languages and anthropology, Aztec, Epi-Olmec, Harrapan, Mayan and Zapotec scripts, typology of logo-syllabic writing systems, iconography and Mesoamerican historical linguistics. He is the founder of the projects on proto-Totonacan (University of Mexico, 2007) and on Aztec script (Bonn University, 2007-2009). Albert carried out linguistic and ethnographic fieldwork with Tepehua de Pisaflores (Mexico, 2007), Sym Evenki and Kellog Ket (Siberia, 2009). E-mail address: aldavletshin@mail.ru Word-signs and sign groups in the Kohau Rongorongo script of Easter Island Russian State University for the Humanities, Institute for Oriental and Classical Studies (Moscow) When confronted with an unknown sign, the first step toward its decipherment should be a consideration of whether it works as a logograph or a syllable. (Stuart 1995: 47-48) Introduction Many papers have been dedicated to the study of the indigenous Easter Island script, called kohau rongorongo. Nevertheless, it remains undeciphered, although the total length of the surviving rongorongo texts (around 11300 glyphs)[1] implies its possible decipherment. Historiographical analysis is beyond this paper (see e.g. Fischer 1997). Moreover, I believe that such work should be written after satisfactory results are achieved; I will mention only the most important -in my opinion- works without commenting on them (Jaussen 1893; Harrison 1874; Piotrowski 1925; Ross 1940; Métraux 1940: 389-411; Kudrjavtsev 1949; Olderogge 1949; Butinov and Knorozov 1956; Barthel 1958; Guy 1982; Pozdniakov 1996; Melka 2008; Horley 2007, 2009). Documentation of the kohau rongorongo inscriptions must be considered good enough and it constantly improves through the joint efforts of scholars (Piotrowski 1925; Ross 1940; Olderogge 1949; Barthel 1958; Fischer 1997; Horley 2009, 2010, 2011). Three points are worthy to note here. Firstly, works dedicated to the Easter Island script usually lack comparative and typological perspectives (for exceptions see Butinov and Knorozov 1956; Kondratov 1969). Secondly, statistical methods and structural analysis of text are popular, whereas combinatory approach, aimed at the study of individual signs, their behaviour and properties, is underrepresented in the field. The combinatory approach has been proven useful in the study of ancient scripts and belongs along with structural analysis of text to conservative methods of decipherment. This situation is surprising, because parallel texts and passages abound in the corpus (Kudrjavtsev 1949; Butinov and Knorozov 1956; Barthel 1958: 151-157) and form a solid basis for such studies. Thirdly, influential scholars are adherent to the idea that kohau rongorongo is essentially a syllabic system (Macri 1996; Pozdniakov 1996; Horley 2005; Pozdniakov and Pozdniakov 2007). This assumption is rather problematic from a typological point of view, because purely syllabic writing systems are rare over the world and their origin owes to the borrowing of the idea of writing from neighbours (see e.g. Gelb 1963; Daniels and Bright 1996). Many scholars assume (e.g. Butinov and Knorozov 1956; Kondratov 1969: 169-192; Fedorova 1982; Guy 2006) that kohau rongorongo is a logosyllabic writing system, a type of script also known as hieroglyphic or mixed. Irina Fedorova (1963: 86; see also Kondratov: 1969) was the first to present evidence favouring the assumption; the number of signs in the script (estimated as some 300 or more signs) is too big for a syllabic script and too small for a purely logographic script; rates for appearance of new signs in a text (counts of new signs found in every 100 signs of a text) are also characteristic of logosyllabic scripts. These arguments, as good as they are, suffer from the fact that hitherto there are no efficient sign catalogues based on a thorough analysis of contexts. Nonetheless, scholars would agree that the inventory of rongorongo signs considerably exceeds the 54 syllables of Rapanui language, if iconic integrity of signs is not violated (cf. Pozdniakov and Pozdniakov 2007). Considering the situation, it is reasonable to undertake a quest for word-signs in the kohau rongorongo script. It is understandable that scholars mostly look for syllabic signs, for they form the basis of any decipherment. Nevertheless, word-signs are an integral part of logosyllabic scripts and might contribute to a decipherment in many different ways. First, word-signs optionally take phonetic complements and can be substituted by a combination of phonetic signs that spell the word. Consequently, they constitute a potential source of phonetic values for syllabic signs. Second, in pictorial scripts, the meaning of the word spelt by a word-sign as a rule is related to the image the word-sign depicts. Pictorial scripts are scripts, whose signs represent recognisable objects, human beings and their actions, for example, Egyptian and Luwian hieroglyphics. They contrast with lineal scripts, for example, Cuneiform, Devanagari and Latin; the last ones make use of abstract graphic designs, that is, combinations of dots, wedges, strokes, etc. Last, but not least, word-signs are not phonetic signs, and these two types of signs shouldn't be confused. The goal of the paper is twofold: to show existing word-signs on the basis of text internal evidence, that is, referring to combinatorial properties of signs only, and to formulate methods that permit to identify particular signs belonging to the class of word-signs. Since I find published sign catalogues (Barthel 1958; Pozdniakov and Pozdniakov 2007) unsatisfactory and misleading, although they give interesting insights, I will use descriptive nicknames for signs such as, for example, "Turtle". Similarly, readers need to consult the figures to identify the signs in question. A nickname doesn't imply a reading value. Its only significance is that the sign looks in a certain way, although I am inclined to use nicknames consistent with my iconographic analysis of signs, if possible. For more clarity, I give a number referring to the generally accepted system of transcription (Barthel 1958), when the sign is mentioned for the first time, e.g. "Turtle" 280. Sometimes this reference is inaccurate, because I differently define a graphic design, for example, "Standing Man with Hair" 200?. Some thoughts on word-signs All well-documented hieroglyphic writing systems possess at least two functional types of signs: phonetic signs (signs that transfer abstract sequences of sounds) and word-signs (signs that spell a word and indicate its meaning), and use the former as terminal phonetic complements in order to clarify the reading of the latter, indicating and at least partially reiterating the reading of word-signs (cf. Gelb 1963). Word-signs frequently spell the stem of the word only, that is, the bearer of its lexical meaning. Specifically, in Maya script a human head in royal headband has the value ʻAJAW LORD and frequently is accompanied by a phonetic sign wa used as a phonetic complement. ʻAJAW alone or a combination of ʻAJAW and wa spell the word ʻajaw "lord". The same word can be spelt with three phonetic signs: ʻa-ja-wa. Semantic determinatives - signs used to indicate the semantic class to which a spelt word belongs are also attested in many hieroglyphic writing systems. Combination of phonetic signs, word-signs and semantic determinatives (if they are an integral part of the system) permits to write any word and phrase in the language for which a particular script has been designed. Word-signs are also called logographs or logograms, and phonetic signs syllabic, hence hieroglyphic systems are defined as logosyllabic. Recognition of different types of signs and rules of their combination has been crucial in the decipherment of writing systems. The popular opinion is that word-signs are unimportant and relatively infrequent in texts. Nevertheless, word-signs refer to the most important cultural concepts and are not less frequent than phonetic signs. In order to illustrate this assertion I have chosen a text written with Maya hieroglyphs (Table). The situation is similar in other logosyllabic scripts. Many word-signs are more frequent than the majority of phonetic signs. Being extremely rare, some phonetic signs are attested a couple of times only in the large corpus of Mayan inscriptions and some haven't yet been found. As a rule, a phonetic sign, whose reading value coincides with a frequent grammatical marker, is the most frequent sign. In Maya writing, it is ʻu mostly found used to spell the third person ergative pronoun and ya used to write the deictic enclitic -iiy. Textually such signs behave as word-signs because they are mostly used to transmit a certain grammatical meaning. Earlier (Davletshin 2002) I have proposed that there are two types of signs in kohau rongorongo texts according to their combinatorial properties. Signs of the first type form sequences of the kind ABAB, AAAA and AAA in combinations with other signs of the same type. Signs of the second type may not appear in such combinations, even if frequently attested in the script. These combinations probably correspond to reduplicated forms typical for Polynesian languages, e.g. Rapanui tea-tea "white", te-tea "whitish", cf. tea "early dawn" (Du Feu 1996: 191). I have also suggested that the ability of a sign to form sequences ABAB and AAAA indicates that the sign has a syllabic (phonetic) value, because such word combinations as "fish fish fish fish" don't make sense in any human language. Consequently, frequent signs which are not attested in these combinations are likely to be word-signs. For technical purposes we can reformulate this observation as Rule 0: if a sign doesn't form ABAB and AAAA sequences and it is a frequent sign, it is a word-sign. However, we need independent combinatorial evidence. The rule for identifying phonetic signs was based on the known observation that reduplicated signs in Maya script are always phonetic. Unfortunately, I was unable to find secure recipes for identifying word-signs. I quote: "If a sign tends to be associated directly with other syllables, then there is a good chance it too is a syllable. This is especially true if the associated syllables all have the same vowel ... Logographs, by contrast, are less likely to be so consistently associated with syllabic signs ... The identification of logographs rests on an assumption that they are signs less consistently associated with syllables and often paired with other logographs" (Stuart 1995: 4849). This difference in behaviour of signs results from their different functions; word-signs spell words and possess meanings on their own, phonetic signs do not. Thus, we can propose Rule 1: if a sign is used mainly on its own (in isolation), it is a word-sign. Importantly, I propose that in many cases ligatures of two or more rongorongo signs include signs used independently. The distribution of word-signs depends on the context more than the distribution of phonetic signs in the same way as the distribution of words in a human language does in comparison with syllables whose distribution is more even. Hence one can formulate Rule 2: if a sign is attested in the same context, in texts of a certain content or genre, it is a word-sign. It should be noted that Rules 1 and 2 are violated in the case, when a phonetic sign is used to spell a monosyllabic morpheme, something that can be called logographic use of phonetic signs (Stuart 1995: 48). In such a case, a phonetic sign behaves like a word-sign. A few lexical morphemes of Rapanui are monosyllabic, but even the use of grammatical morphemes heavily depends on the context and genre. The remedy in such situation is obvious – frequent phonetic signs are expected to be found in other contexts, in particular, being constituents of sign groups used to write words. Last but not least, word-signs can take phonetic complements and can be substituted by two or three phonetic signs. This brings us to Rule 3: if a sign is substituted by two or three signs in the same context, it is a word-sign. The definitions suggest that word-signs and sign groups are similar in their behaviour, because in contrast to single phonetic signs they have semantic value. Such sign groups are expected to be combinations of phonetic signs used to write words. Methods of detecting word-signs in the Kohau Rongorongo Script The "Chief" sign 530 (Davletshin 2002) frequently appears on the Santiago Staff (93 times) and on the Honolulu Tablet B.3629 (5 times). At the same time it is rare in the other texts, with only few instances attested (e.g. Tahua and Mamari Tablets). As I have suggested previously, Santiago Staff and Honolulu Tablet probably record lists of personal names. "Chief" is mostly attested in the same position, at the beginning of the name after the "Kohe-plant" sign 076 (85 of 98 examples). These observations indicate that the sign in question is a word-sign for a title, possibly intended to be read "chief" ʻARIKI?. Its distribution depends on particular texts, their genre and "Kohe-plant", suggesting a logographic value (Rule 1). It is used alone and it is difficult to imagine this sign having a syllabic reading (Rule 3). The sign in question is not attested as a member of ABAB and AAAA sequences (Rule 0). Looking at various versions of the sign (Fig. 1a), we might see that it depicts a "man" with a "feather headdress", sometimes, bearing "feather standards(?)" behind his shoulders and one or two "spherical objects" attached to his arms with a string (tahonga wooden pendants?). The "man" depicted seems "to touch the ground with his hands" and sometimes "one of his hands is raised" (dancing?)[2]. Remember that Polynesians also dance sitting. Such an image is an apt representation for a chief. Two other possible titles attested on the Santiago Staff and Honolulu Tablet and similar in their behaviour are sign groups (Fig. 1a'). Examining parallel passages of rongorongo texts, which were originally discovered by Boris Kudrjavtsev (1949), we can see that there are cases when a sign is relocated in parallel texts. Such transpositions are of two kinds: close –when two signs swap around -and distant – when a sign is relocated several signs forth or back (Fig. 2a, b). Distant transpositions as we can see in case of "Sea Creature" 755 (Fig. 2b) are good evidence for a logographic value, for word order changes while rewording, but it is difficult to imagine why a syllable would change its place when the same phrase were said in a slightly different way. Sign groups are also amenable to transpositions (Fig. 2a', b'). Examples of isolated signs may be seen in cases of palindromes (Fig. 2c), that is, mirrored sequences of signs (Barthel 1958: 164), and sign lists (Butinov and Knorozov 1956: 81-82), when the invariable part of an enumeration is a single sign (Fig. 2d). Examples of the last type are not so secure, because we can suspect logographic use of phonetic signs in lists. A good example of possible logographic use in kohau rongorongo is "Kohe-plant" 076, which functions as a marker of personal names in the lists attested on the Small Santiago Tablet Verso, Honolulu Tablet B.3629 and Santiago Staff, although its identification as a phonetic sign heavily depends on the external data, namely, name lists in Rapanui texts (Davletshin 2002). Lists and palindromes also include sign groups (Fig. 2b', 2c'). Word-signs in the Kohau Rongorongo Script There are extremely formalized text passages in rongorongo texts which unambiguously demonstrate that some signs are used alone and not in combination with others (Fig. 3). One of them (Fig. 3a) is the genealogy on the Small Santiago Tablet discovered by Yuri Knorozov and Nikolai Butinov (1956: 87, 89). The structure of the passage (Davletshin 2002) might be described as the following: -A-B-X i -X i+1 , ... where A is "Kohe-plant" 076, B – "Standing Man with Hair" 200?, and X – a variable consisting of one or two signs. Another one (Fig. 3b) represent three sequences of signs found on the Aruku Kurenga Tablet (Butinov and Knorozov 1956: 82-83): -A-X i , -B-X i , -C-X i , where A is "Crippled Man" 230 of Line 3, B – "Sitting Man" 380 of Lines 3 and 4, C – "Adze" 063 of Line 6 and X - a variable consisting in one or more signs attested in every sequence. Yuri Knorozov and Nikolai Butinov (1956: 82-83) write: "Combinations ... are of particular interest. In them one can distinguish separate words". It does not matter how these texts should be interpreted: when X is a single sign, it is clearly used alone. This suggests a logographic value, and such examples are paired with cases where X is a sign group. Thanks to the two text fragments, the following signs can be identified as word-signs: "Turtle" 280, "Shark" 720-721, "Sea Creature" 755, "Knife" 004, etc. The same reasoning can be applied to numerous sign lists attested in the kohau rongorongo texts (Butinov and Knorozov 1956), where a variable part of an enumeration is a single sign. Taking into account that lexical monosyllabic morphemes are uncommon in Rapanui, it is impossible to imagine that all these examples represent phonetic signs in their logographic use. A list of signs depicting "plants" is attested on the Great Santiago Tablet (Fig. 4a). Each unit of the list is introduced with "Twig" 068; the entries always consist of a single sign: -A-X i , where A is "Twig" and X is a sign depicting a "plant". In the nearby passage there is another list of signs depicting "plants" (Fig. 4b): -B-X i , where B is "Crescent" 040-042 and X is a plant-like sign. The X signs are indicative of word-signs, because they are found alone (Rule 1) and depend on the context (Rule 2). Assuming they are syllabic, we need to admit that every entry in the lists possesses its own meaning, corresponding to a word and spelt by one syllable only; the conclusion drawn can not be correct, for monosyllabic words are very rare in Rapanui. Some logosyllabic scripts don't like to write the same phonetic signs twice; for example, in Aztec script we:we:-tl "slit-drum" is to be written as we, but not as we-we. It is not the case for the kohau rongorongo script, for the AA combinations abound in texts. There is one more argument in favour of their identification as word-signs: all the X signs depict "plants" – "Two Leaved Stalks" 031?, "Berried Plant" 124, "Fern Sprout" 059, "Leaved Sprout" 048-049, "Stalk" 011?, "Hanging Fruit" 074, "Sprout" 062?. It is worth saying that exact iconographic analysis of these particular signs and related nicknames are by no means certain. Nevertheless, the discussed signs look like plants and share iconic designs with each other and with other signs depicting plants, e.g. "leaves" ("Two Leaved Stalks", "Fern Sprout", "Leaved Sprout", "Hanging Fruit"), "berries" ("Berried Plant") and "roots" ("Two Leaved Stalks", "Berried Plant", "Fern Sprout"). In my opinion, these lists represent the best evidence for existence of word-signs in the kohau rongorongo script, for it is impossible to imagine such a structured sequence of signs depicting homogeneous objects occurring purely by chance, while the meaning of the word spelt by a word-sign is frequently related to its shape. Interestingly, similar lists of plants are also attested in Rapanui folklore texts (cf. Manuscript E69, Barthel 1974: 359-360). I identified two signs, systematically substituted by sign groups of the type "the same sign plus one more sign" in several contexts (Fig. 5): "Crescent" 040-041 is replaced by combination of "Crescent" and "Standing Man" 200 and "Adze" 063 - by "Adze" and "Poker"(?). Such behaviour of the signs "Standing Man" and "Poker" is typical for syllabic signs used as phonetic complements. Thus, "Crescent" and "Adze" are to be word-signs. The best example for an isolated rongorongo sign may be seen between the symbols carved on heads of wooden figurines, which I consider to represent a decorative font of the kohau rongorongo script. Again here isolated signs (word-signs) go with sign groups (see e.g. Orliac and Orliac 2008: Fig. 64). A challenge The reader can notice that, at least in a few cases, there is a disagreement between combinatorial properties of signs. "Knife" 004 is a word-sign according to its use in isolation, but it is also attested as a part of AAA, ABAB and BABA combinations, which suggest its phonetic value. The same is true for "Sea Creature" 755. One can interpret it as a downside of the methods proposed here. However, this disagreement might result from a lack of combinatorial data at our disposal. It is probably the case of "Sea Creature", because it occurs in only one ABAB combination and this combination is part of a longer structure ABABC(DE)AF, where A is "Sea Creature" (Great St.Petersburg Tablet Verso line 6 and parallel texts). In fact the ABABC(DE)AF structure suggests a logographic value of the sign. Two other functional explanations for the disagreement in question are possible. First, phonetic signs could have logographic uses. Second, signs could have multiple reading values; homographic signs are common in some logosyllabic scripts, for example, in the Sumerian script and Aztec. In order to choose among these two options a thorough analysis of the combinatorial properties of as many rongorongo signs as possible is necessary. Issue of semantic determinatives Semantic determinatives, which are defined as signs that have no phonetic value, but indicate the semantic class to which a spelt word belongs, are common in many logosyllabic writing systems. They are situated either at the beginning or at the end of sign groups depending on a particular writing system and sometimes replaced by other determinatives, but almost never omitted (cf. Fischer 1995; Guy 2006: 66). They cover the majority of lexical items attested in a script. Scholars have proposed semantic determinatives for kohau rongorongo (Butinov and Knorozov 1956: 86; Fischer 1995: 313, 316; Guy 1998: 554; 2006: 65, 66; Horley 2007: 27; Melka 2009: 44). Looking at the list of plants examined above (Fig. 4a, cf. Fig. 2d, Fig. 4b), one is eager to suggest that "Twig" behaves as a determinative for plants. However, as shown in this paper many rongorongo signs are used alone and it implies that determinatives are not a part of the system, at least, that they do not form an extensive class of signs. In contrast to word-signs, determinatives in some scripts such as Luwian constitute a highly restricted class that consists of only a few signs. Another reasoning may be applied to corroborate this observation. Let us assume that "Twig" is a semantic determinative for plants and "Kohe-plant" is a determinative for personal names as their lists would suggest. Then we would find that personal names and plants are not found in the rongorongo texts beyond the discussed lists. The corollary we have arrived at cannot be correct for it is impossible to imagine that almost all the texts lack personal names and we find signs depicting "plants" all over the texts. I was unable to detect rongorongo signs, whose behaviour is indicative of semantic determinatives according to the stated above definitions. None of the numerous pre-Columbian writings of Mesoamerica have been proven to possess semantic determinatives, and two of them - the Maya script and the Aztec one - are well understood today. Therefore, the conclusion that kohau rongorongo lacks semantic determinatives is typologically grounded. Conclusions It has been possible to propose evidence for the existence of word-signs in the kohau rongorongo script by referring to their combinatorial properties. My intention is not to supply an exhaustive list of word-signs, but rather to devise methods that permit detecting them. The reader familiar with rongorongo tablets will easily remember many examples of transpositions, sign lists and other cases of signs used in isolation. Signs "Adze", "Berried Plant", "Chief", "Crescent", "Fern Sprout", "Hanging Fruit", "Leaved Sprout", "Nightshade", "Sea Creature", "Shark", "Turtle", "Twig" and "Two Leaved Stalks" should be considered as good candidates to be read as words and not as syllables. Expectedly, some sign groups behave as wordsigns. These might refer to words, and their constituents probably represent phonetic signs. My estimation is that kohau rongorongo is a heavily logographic script with a robust syllabic component. The methods for detecting wordsigns and phonetic signs should be carefully applied to every single sign of the script and the results of such application should be constantly checked and revised. Acknowledgements I would like to thank Martyn Harris for inviting me to participate in the conference "Easter Island: Cultural and Historical Perspectives". Thanks to him my stay in London was great and fruitful. I am also very grateful to many friends and colleagues who have helped through discussion and providing sought-for materials: Artem Kozmin, Dmitri Beliaev, Paul Horley, Rafal Wieczorek, Scott Nicolay and Tomi Melka. I am very much obliged to Paul Horley for his kind permission to use his unpublished tracings of rongorongo glyphs and to Evgenia Korovina who efficiently helped me to find an illustrative example (Fig. 2c'). Many thanks go to the curators Pavel Belkov and Tatiana Sokolova (Peter the Great Museum of Anthropology and Ethnography, St.-Petersburg) and to Jill Hasell (British Museum, London) for the opportunity to work with the original tablets. This study is based on results obtained during my six-month research stay at the Department Anthropology of the Americas, Bonn University, supported by an "Immanuel Kant" scholarship (2009-2010). References: Barthel, Thomas S. (1958): Grundlagen zur Entzifferung der Osterinselschrift. Abhandlungen aus dem Gebiet der Auslandskunde 64, Reihe B: Völkerkunde, Kulturgeschichte und Sprachen Band 36. Hamburg: Cram, de Gruyter. Barthel, Thomas S. (1974): Das achte Land: die Entdeckung und Besiedlung der Osterinsel. München: Klaus Renner. Butinov, Nikolai A., and Yuri V. Knorozov [ . ., К Ю. .] (1956): э я 4: 77-91, see also translation: . С е я Butinov, Nikolai A., and Yuri V. Knorozov (1957): Preliminary report on the study of the written language of Easter Island. Journal of the Polynesian Society 66/1: 517. Daniels, Peter T., and William Bright (eds.) (1996): The world's writing systems. New York-Oxford: Oxford University Press. Davletshin, Albert (2002): Names in the Kohau Rongorongo script. Paper presented at the 2nd International Conference "Hierarchy and Power in the History of Civilizations", St.-Petersburg, July 4-7, 2002. Du Feu, Veronica M. (1996) Rapanui. Descriptive Grammars Series. London-New York: Routledge. Fedorova, Irina K. [Фё К С е я э .К.] (1963): я я 2: 85–92. . Fedorova, Irina K. [Фё .К.] (1982): . : ( ). : : Ю. . К ( , я , : 23-98. .): я Fischer, Steven R. (1995): Preliminary evidence for cosmogonic texts in Rapanui's rongorongo inscriptions. Journal of the Polynesian Society 104: 303–321. Fischer, Steven R. (1997): Rongorongo, the Easter Island script: History, traditions, texts. Oxford Studies in Anthropological Linguistics 14. Oxford and New York: Oxford University Press. Gelb, Ignace J. (1963): A study of writing: a discussion of the general principles governing the use and evolution of writing. 2nd Edition. Chicago-London: University of Chicago Press. Guy, Jacques B.M. (1982): Fused glyphs in the Easter Island script. Journal of the Polynesian Society 91: 445-447. Guy, Jacques B.M. (1998): Easter Island – Does the Santiago Staff bear a cosmogonic text? Anthropos 93/4–6: 552–555. Guy, Jacques B.M. (2006): General properties of the rongorongo writing. Rapa Nui Journal 20/1: 53-66. Harrison, J. Park (1874): The hieroglyphics of Easter Island. Journal of the Royal Anthropological Institute of Great Britain and Ireland 3: 370-382. Horley, Paul (2005): Allographic variations and statistical analysis of the Rongorongo script. Rapa Nui Journal 19/2: 107–116. Horley, Paul (2007): Structural analysis of Rongorongo inscriptions. Rapa Nui Journal 21: 25-32. Horley, Paul (2009): Rongorongo script: Carving techniques and scribal corrections. Journal de la Société des Océanistes 129/2: 249-261. Horley, Paul (2010): Rongorongo Tablet Keiti. Rapa Nui Journal 24/1: 45-56. Horley, Paul (2011): Palaeographic analysis of the Santiago staff. Rapa Nui Journal 25/1: 31-43. Jaussen, Florentin Étienne (1893): L'île de Pâques: historique – écriture, et répertoire des signes des tablettes ou bois d'hibiscus intellegents. Bulletin de la Société de la Géographie 2: 240-270. Kondratov, Alexander M. [К C . я. . . Kudrjavtsev, Boris G. [К . С э . .] (1969): . ё , я . .] (1949): ея o 11: 175-221. Macri, Martha J. (1996): Rongorongo of Easter Island. In: Peter T. Daniels and William Bright (eds.): The world's writing systems. New York-Oxford: Oxford University Press: 183–188. Melka, Tomi S. (2008): Structural observations regarding rongorongo tablet 'Keiti'. Cryptologia 32: 155-179. Melka, Tomi S. (2009): Some considerations about the Kohau Rongorongo script in the light of a statistical analysis of the Santiago Staff. Cryptologia 33/1: 24-73. Métraux, Alfred (1940): Ethnology of Easter Island. Bernice P. Bishop Museum Bulletin 160. Honolulu. Olderrogge, Dmitri A. [ ". С ея . .] (1949): o "К э 11: 222-236. Orliac, Catherine, and Michel Orliac (2008): Trésors de l'Île de Pâques / Treasures of Easter Island. Paris: Éditions D / Éditions Louise Leiris. Piotrowski, A. (1925): Deux tablettes avec les marques gravées de L'île de Pâques, de la collection de N. N. Mikloukho-Maklay. Revue d'Ethnographie et des Traditions Populaires 6/23-24: 42531. Pozdniakov, Igor, and Konstantin Pozdniakov (2007): Rapanui writing and the Rapanui language: preliminary results of a statistical analysis. Forum for Anthropology and Culture 3: 73-36. Pozdniakov, Konstantin (1996): Les bases du déchiffrement de l'écriture de l'île de Pâques. Journal de la Société des Océanistes 103/2: 289303. Ross, Alan S.C. (1940): The Easter Island Tablet Atua-mata-riri. Journal of the Polynesian Society 49: 556-563. Stuart, David (1995): A study of Maya inscriptions. Ph.D. dissertation. Vanderbuilt University, Department of Anthropology: Nashville. Footnotes: [1] There is no consensus on how to count rongorongo signs, so the given numbers are obtained by counting writing units separated by blank spaces according to Thomas Barthel's transcriptions (1958: 15-33; the reader should sum up numbers for individual texts). Many of such units represent ligatures of signs, but this is a useful and fast method to estimate the size of the corpus with our incomplete understanding of kohau rongorongo graphics. Documentation has been improved since 1958, so that the total length of the corpus is larger. [2] Paul Horley (pers. commun. 2011) suggested to me that "Chief" depicts a "man touching the ground with his hands" and the "balls above his shoulders" are tahonga pendants. Table. Twenty most frequent signs on the Tablet of 96 Hieroglyphs from Palenque. Word-signs are given in capitals and supplied with translations, phonetic signs - in small letters, diacritical marks – in superscript. Slash "/" indicates multiple values of signs, numbers are used to distinguish allographs (different signs with the same reading value). Total of signs in the text is 356. Reading values of signs: Number of signs in the text: 1. ya 20 2. la 11 3. ʻu 10 4. ji 10 5. K'IHNICH[INCANDESCENT] 8 6. le 8 7. ʻAJ[REED?] 7 8. 9. ʻAJAW[LORD] DAY 10. ʻAJAW[LORD] 2 7 7 6 11. ʻi 6 12. BAAK[BONE] 6 13. ti 6 14. WINAAKHAAʻB[TWENTY_YEARS] 6 15. ku/TUUN[STONE] 5 16. K'UH[GOD] 5 17. NAAH[HOUSE] 5 18. wa 5 19. BOLOʻN[NINE] 2 4 20. CHAʻ[TWO] 2 4 List of figures: Figure 1. Possible titles in kohau rongorongo: A) different versions of the "Chief" sign on the Santiago Staff; A') other possible titles. After Paul Horley's drawings by his courtesy. Figure 2. Word-signs in kohau rongorongo: A) close transpositions, B) distant transpositions, C) palindrome sequences of signs. D) List introduced with the "Nightshade" sign 035, 125. Texts A, B, E, P and Q: after Paul Horley's drawings by his courtesy. Hr2: after drawings by Bodo Spranz (Barthel 1958). Hr11-12: after drawings by Steven Fischer (1997). From here on, the arrows indicate probable word-signs, numbers are used to show identical word-sings and numbers prefixed with "s" refer to sign groups. Figure 3. Word-signs and sign groups in kohau rongorongo: A) the so-called genealogy on the Small Santiago Tablet; B) the twice transformed sequence on the Aruku Kurenga Tablet. After Paul Horley's drawings by his courtesy. Figure 4. Lists of signs depicting "plants" on the Great Santiago Tablet. After Paul Horley's drawings by his courtesy. Figure 5. Phonetic complements in kohau rongorongo: A) "Crescent" and its phonetic complements, B) "Adze" and its phonetic complements. Text B, D, H and P: After Paul Horley's drawings by his courtesy. Text E: drawings by the author after photographs from Horley 2010. Asterisks indicate probable phonetic complements. Figure 1. Figure 2. Figure 3. Figure 4. Figure 5.