Navigation – Plan du site

From icon to sign

Local iconography and the birth of the cretan hieroglyphic script
Silvia Ferrara
Traduction(s) :
De la figure à l’écriture


This article offers a reconstruction of early language recording on Crete in the second millennium BCE, charting the development of the earliest attestations of script on the island and examining their relationship to elements of iconology, orderly codification, figurative representations and symbolic expression. From an interdisciplinary perspective, it is possible to reconstruct an autonomous path into script formation starting from an analysis of the designs on Prepalatial, but especially, Protopalatial Cretan seals (from the third millennium BCE to the beginning of the second). Cretan hieroglyphic is still an undeciphered script. Yet, despite this hurdle, it is possible to recognise the stage in which phonetic recording takes place, through an in-depth analysis of the archaeological record and the inscribed objects, and through an in-depth examination of the structural/semantic categories represented on the inscriptions. On this basis, it is possible to conclude that the earliest Cretan language, whatever it may be, harnesses existing iconography on seals to be part of the inventory of its signs.

Haut de page

Entrées d’index

Haut de page

Texte intégral

Amulette sind dergleichen / Auf Papier geschrieben Zeichen.
“Les amulettes sont des signes du même genre, tracés sur le papier.” Johann Wolfgang von Goethe, « Segenspfänder » (1861 : 533).

  • 1 I am very grateful to John Bennet for having commented on this paper, and for his constant and unf (...)

1Reconstructing the setting for the earliest instances in which language is recorded for the first time on Crete is particularly challenging, not least because its earliest form of writing is still undeciphered1. The aim of this chapter is to set the scene for this complex phenomenon from a polycentric perspective that revolves around archaeology, art, language, and cognitive perception. More precisely, it will focus first on (1) an analysis of the archaeological context in which the (purported) earliest attestations of Cretan writing were found, (2) considering the iconography and symbols that may have provided the basis for the Cretan writing, (3) investigating early phonetic coding, considering possible semantic categories used on the stone seals, such as logograms and determinatives, (4) analysing the features of the most frequently attested and often repeated sequences, the so-called formulas, on sealstones from the period of the first palaces (called conventionally ‘Protopalatial’), and (5) finally proposing a pointed birth of the fully-developed writing system during the MM II period (approximately from 1850 BCE), after a long and gradual development.

The earliest writing on Crete?

Fig. 1. Phourni Cemetery Site Map, Archanes, Crete

Fig. 1. Phourni Cemetery Site Map, Archanes, Crete

Croquis : Silvia Ferrara

2Phourni is a necropolis located in the north-central portion of Crete, close to the village of Archanes. It is one of the most complex cemeteries on the island, comprising different architectural typologies of tombs and a vast array of funerary assemblages. Its occupation is especially important for the phases that precede the emergence of the palaces, as it reached its peak at the transition between the third and second millennium BCE. With exceptions, such as Tholoi B and Gamma, Burial Building 19 and the deposit outside Burial Building 6, a large part of the material found at the cemetery has not been published in detail, and a thorough investigation of the exact find-spots of the seals cannot be carried out. From a close analysis of the assemblages that are available, it appears that the non-ceramic material is quite uniform in typologies (Legarra Herrero 2014: 82 & fig. 56D), showing that the cemetery is particularly rich in seals, and contains numerous examples of metal objects and stone vessels.

3Burial Building 6 is important because it is in its confines that the earliest “attestations” of writing on the island were recovered and ascribed to the crucial transition from Prepalatial to Protopalatial. With regards to the details of this material, the main publication of the site states that seven seals bearing hieroglyphs were discovered, four in Burial Building 6, two in Burial Building 3, and one in Burial Building 7 (Sakellarakis & Sapouna-Sakellaraki 1997: 327). The absence of a detailed publication of the whole site hinders an examination of the seals recovered in the latter two buildings. As a result, they are not included in the main corpus of Cretan hieroglyphic inscriptions by Olivier and Godart, the Corpus Hieroglyphicarum Inscriptionum Cretae, known by the acronym CHIC.

  • 2 Olivier and Godart (1996) point out that the specimens from Samothrace, Knossos and Gouves only be (...)

4The remaining four objects, all from Burial Building 6 (fig. 2) show incised symbols that have been linked to writing, albeit in a sui generis form, which has been singled out from the other two scripts of Protopalatial Crete, and termed “Archanes script” (Yule 1980). The four seals from Building 6 all share an identical “formula,” which is also found on other seals from other sites on the island and beyond, as far as Samothrace (CHIC #134-#137, #179, #201, #205, #203, #292, #313; Schoep 2006: 45).2 It also appears, crucially with some variations, on votive vessels inscribed in Linear A from the Neopalatial period (MM III-LM IB, approximately 1700-1450 BCE) (Godart 1999; Anastasiadou 2016).

Fig. 2. Attested inscriptions in the cemetery of Archanes-Phourni

Corpus der Minoischen und Mykenischen Siegel (CMS). CHIC Museum inventory Context within Archanes Typology and material Inscription
II.1, 391 #315 HM 2260 Burial Building 6, Room III 14-face seal, bone “Archanes formula” + isolated logograms?
II.1, 393 #252 HM 2266 Burial Building 6, Room III Gable seal, bone “Archanes formula”
II.1, 394 #202 HM 2245 Burial Building 6, Room III Discoid seal, bone “Archanes formula”
VI.I, 14 #251 AM 1938.928 From village of Phourni Gable seal, green steatite “Archanes formula?”
  • 3 Recueil des inscriptions en linéaire A, Godart & Olivier 1976-1985.

5As convention dictates, an inscription is defined as the association of at least two, or more, consecutive signs arranged in a coherent sequence, with a consistent orientation (GORILA3). The examples that appear at Archanes-Phourni are engraved on a series of seals, one being a fourteen-face object, a typology which has no parallels elsewhere (CHIC #315, fig. 3). On this specimen, the Archanes formula is engraved on two adjoining faces, which were likely used as a matrix for stamping (see faces h and i). As for the other faces, it is impossible to ascertain whether they were also inscribed with hieroglyphs. The fact that the inscriptions come from the same context and show epigraphic uniformity is in itself testament to the likelihood that this formula was associated with a specific clan or family, possibly related, who made a point of using similar shapes and motifs (Karytinos 1998). These objects orbited around the central portion of the site and were connected to some high status individuals or groups.

Fig. 3. Faces of Archanes’ seal

  • 4 The Cretans may well have had awareness of Egyptian literacy, but we are in the dark as to whether (...)

6This is a particularly meagre sample, whose nature is repetitive and the repertoire of signs is so limited, to such an extent that we cannot define this as a self-contained, complete system of writing. Despite this, we find several clues, specifically in the cemetery of Phourni, that point to a phenomenon involving what we may define as “iconic literacy” at play; a thin shadow line that signposts a complex interplay between iconography and language coding. The pre-existing literate culture of Egypt may have seeped through Crete and may have played a part in introducing visual narratives that were conducive to linguistic representation. Yet, no specific Egyptian hieroglyphic model can be detected, nor should it be looked at as an explanation for the formative stages of Cretan writing.4 And despite the fact that Arthur Evans classed Cretan signs as “blundered imitations, travesties or imperfect renderings of Egyptian inscriptions” (Evans 1894: 316), it can now be concluded that the similarity between the two scripts is but arbitrary, the name itself of ‘hieroglyphic’ misleadingly betraying the fact that both scripts rely on iconicity and have a strong pictographic quality in the rendering of their sign shapes.

  • 5 Cat. no. 50 in Phillips 2008: HM 464. Cf. also CSM II.1: 468, no. 395.
  • 6 In the same building, a bone seal, in the shape of a fly, seems to be slightly earlier (EM III-MM (...)

7Egyptian or Egyptianising scarabs are, in any case, present on Crete and in Archanes-Phourni specifically (three on the whole site, according to Phillips 2008: 30-42; two according to Sakellarakis & Sapouna-Sakellaraki 1997: 357), together with other such objects, such as stone vessels, a sistrum, amulets and pendants. It is particularly significant that the earliest imported scarab on the island, dating to the beginning of the second millennium BCE, is found in Burial Building 6.5 Engraved on the base of the scarab are unintelligible motifs, which may have represented imitative patterns based on Egyptian. To the best of my knowledge, this detail has not been discussed in any specificity. It is interesting that this object was found in the same room as the 14-face seal, with most of the other Archanes inscriptions (fig. 2).6

8The symbols on the base of the fly are by no means the only instance of motifs found in the Archanes-Phourni material, which, in some way, can be reminiscent of the Cretan hieroglyphic signary. Others are listed in fig. 4 and all come from Burial Building 6. One of these is a cylinder seal, and bears the engraved symbol of, again, a sistrum. The seal is peculiar, its cylindrical shape is hollow inside and is covered by two thin bone lids, whose surfaces are engraved and secured by rivets. According to the excavator, it was “perhaps designed to contain some magic material” (Sakellarakis & Sapouna-Sakellaraki 1997: 677).

  • 7 Phillips maintains that the typology of sistrum found at Archanes has a closed, arched top found i (...)

9Whatever its function, the representation of the sistrum is not unique to Archanes. 7

Fig. 4. Other possible script signs of the Cretan script attested at Archanes-Phourni with the same chronological horizon (EM III-MM IA)

CMS CHIC Museum Inventory Context within Archanes Typology and Material Inscription
II.1, 379 --- HM 2251 Burial Building 6, Room I Fly-shaped seal, bone Two possible signs on base? 028-066-bifoliate figure<->
II.1, 389 --- HM 2254 Burial Building 6, Room I 3-face seal, black steatite Face c may have signs? No correspondences with the later repertoire
II.1, 392 --- HM 2246 Burial Building 6, Room I Cylinder seal, bone Sistrum: logogram?

10What can this interplay between physical objects and their representation mean in this context? This excursus on the depictions and evidence of sistra at the site brings us back to the instance of its isolated attestation on the 14-face seal from Burial Building 6 (fig. 3, face (d)), and all attestations of symbols that, at this stage, are self-standing and isolated, not juxtaposed to a series of elements. We have stated with due caution that an inscription on this object is borne on two of the 14 faces. The CHIC makes this statement a fact, by excluding all discussion on the other faces, which are seen as carrying elements of iconography, rather than signs, and thus not even drawn in the facsimiles reproduced in the corpus (CHIC, #315: 290). Sakellarakis and Sapouna-Sakellarakis are less assertive in their position and postulate that signs on this seal such as the sistrum, as well as others, may be ‘individual hieroglyphs,’ but they do not venture further into the possibility that these symbols may have a phonological, or logographic, value (Sakellarakis & Sapouna-Sakellaraki: 328).

11On the seal there are single representations of a hand and a leg (fig. 3, faces (j) and (l)), which, together with the sistrum, correspond symmetrically with two signs of the Cretan hieroglyphic signary, as signs 08, 010, and 057 respectively (CHIC, 17; Krzyszkowska 2005: 70-71). Could these be employed in the same way as the script signs? There is no reason to assume that they should not. Moreover, face (k) shows a man holding a basket, and, according to Sakellarakis, the same basket seems to be represented on face (f). By analogy with the sistrum taken to be a logogram, it could be postulated that all the symbols of the second matrix on the seal are logograms, as well as those on the last matrix. In other words, there is a pattern, and the fact that the symbols appear in isolation should not deter from postulating a possible phonetic interpretation. Less probable is a phonetic value for the set of quadrupeds on the other faces, or the rosette motif, which seem to be oddly similar, or too generically decorative, to function as sematograms (although Sakellarakis is less cautious in ruling this out, but see also Krzyszkowska 2005, 71). As becomes immediately apparent, the boundary between art and writing reveals lines which are difficult, but not impossible, to discern.

12The Archanes multi-face seal is the introductory probe into the problem we are facing that needs to be examined further. When does an image become a sign? There are interpretative complications, to be sure. This is particularly true for symbols that appear in isolation. According to the standard definition, these symbols, by virtue of being isolated, cannot be defined as proper inscriptions. Therefore we cannot grasp whether they may be phonetically representative or not, since they lack the necessary context or association with other signs. In more general terms, this very hindrance can be extended to all isolated symbols engraved on the Cretan seal corpus, and as such it represents a crucial aspect that needs further investigation: when and where does language notation on the island first take place? And can we set out possible criteria that enable us to delineate the boundary that separates art from writing, and iconography from language notation?

  • 8 We would like to warn the reader that every letter in blog and in parantheses correspond to a sign (...)

13Crete’s earliest forays into possible “writing” present us with a series of complications. The attestations of the “Archanes formula,” composed of five “signs,” show both symbols with a pronounced level of abstraction, such as sign 019 [h]8, the so-called “cuttlefish,” and figurative icons, such as the double axe (corresponding to sign 042 [ó] in the normalised repertoire, and the jug, corresponding to sign 052 [®]). So, one formula combines both iconic and abstract signs together, to reveal that at this point in time, this “script” is at a level at which signs cease to be sheer representations of iconicity and already proceed into graphic stylisation (fig. 5). I shall not enter the debate over whether this formula is to be considered Cretan hieroglyphic or Linear A (Godart 1999; Karnava 2000; Anastasiadou 2016), as this is a problem of terminology and epigraphic tradition, thus not particularly relevant to the issue at stake here, which is namely whether it records language (whichever language) or not.

Fig. 5. Multifaces seals of Archanes

Fig. 5. Multifaces seals of Archanes

14It has already been pointed out that the Cretans who developed the earliest writing must have been exposed to pre-existing literate environments. However, the signary they design is new and the cognitive processes that underpin phoneticism define this phenomenon as a proper invention rather than a reworked imitation. The possibility that this may be the case should be investigated, as I will show further below. The signary retains a high level of iconicity in the shape of the signs, at least to the extent showcased in the repertoire on the stone seals. This retention is observable well into the MM III period. As for syntactical notation, on the earliest as well as subsequent Cretan seals, it is curtailed to the minimum. In light of these observations, the Cretan hieroglyphic as a whole seems to behave like a pristine writing system, developed independently of a direct influence (Ferrara 2015).

15If we accept that the Cretan script may have developed locally, rather than from external stimulus or derivation, we will need to consider the background that made the formation of a fully-fledged writing system possible. Can we see a substratum, whose characteristics can indicate a propulsive trajectory or a prompt towards the development of writing?

Seal designs as precursors of writing

16In both Mesopotamia and Egypt, the earliest instances of writing, which have been defined “proto-writing” (Damerow 1999; Baines 2004), are not language-bound, and as such they do not bear any relation to the specifics of linguistic notation. This applies also to the earliest stages of Anatolian hieroglyphic (Marazzi 1990) or to pre-Maya writing (Houston 2004), which were, most likely, not intended to record any particular language, but worked as “emblematic” pictograms. Essentially, what the sign represented pictorially was indeed what the sign signified semantically.

17Only to a limited degree has the scholarship on Cretan hieroglyphics considered the early stage of writing formation, which appears on the stone seals, and its relation to the preliterate usage of symbols and imagery on non-inscribed stones. This has engendered a neat, if somewhat artificial, separation of the script from all other iconological features, designs and motifs incised on the seals, severing the two spheres as if they belonged to unrelated fields of analysis. So, any in-depth analysis of the writing system has only taken into account matters that related to writing (Olivier 1989, CHIC), and disconnected it from all other iconographic features that pertain to the non-writing sphere. By the same token, studies of the glyptic imagery are, by their very scope, based principally on style and iconography (Krzyszkowska 2005; Anastasiadou 2011). A joint analysis of the seal designs vis à vis the hieroglyphic signs is still largely a desideratum in the scholarship (but see Decorte 2017).

18There is scope to study the hieroglyphs as a semiotic system at-large, fundamentally tied to a reservoir of symbols that may have less to do with language coding in the strictest sense, than to the conveyance of messages in the same “pictographic” way that proto-cuneiform or proto-Egyptian may have done. How close is a relationship between seal iconography and the proper signary? This question has not received full attention, and calls for a closer probing. The aim should be to focus on the iconography of the seal traditions in the pre-literate stages, to investigate the possibility that these may represent a visual background for the inception of literacy.

19In other areas of the world, like Mesopotamia (Green 1981; Pittman 1994) and Egypt (Baines 2004; Wengrow 2011), the study of proto-literate artistic traditions has revealed a long-standing concern with wider modes of image making, and the possibility that visual representations may have provided a backdrop for symbolic expression that was closely related to, and fashioned in parallel with, writing. An integration between visual imagery and written texts is characteristic of a specific, traditional Egyptian style or decorum (Baines 2004), and the pictographic roots of writing at a stage where it is not yet phonetic can be seen in the tags found in the U-j tombs at Abydos (most recently, Dreyer 2011). The same could be argued for the proto-cuneiform glyptic imagery of the Uruk IV and III periods, in which visual narration, be it presented through simple, schematic emblems or through more complex presentation of figural depictions, partakes of the same orderly arrangement that characterised the inception of writing in the same period (Pittman 1994): the two modes of expression shared features that partook of what could be assumed to be the same cognitive framework; two not-too-separate facets of a larger semiotic system. The fundamental principles that were common to both can be identified in the invariance and standardisation of signs, the legibility of individual symbols or composition, and a tendency for orderly representation. At the earliest stages, in light of the loose dependence on the specificity of language, a pictogram could take the role of determinative, logogram or ideogram, and be combined with other non-linguistic signs, images and forms of communicative actions. Script and image thus emerge out of the same cultural reservoir.

20There is scope to investigate a possible overlap between iconography and proto-script in the Cretan context, too. During the EM II period, ornamental, geometric and abstract motifs such as spirals and circles predominate in the glyptic designs, with various combinations of cross-hatching and schematic shapes populating the seals. Such background presented a more marked stylistic homogeneity, with a by-and-large absence of individualising features (Sbonias 1999; Decorte, 2017).

21During the transition from Prepalatial to Protopalatial (EM III-MM IA), a quantum leap is visible in the glyptic designs, which may bear significant testimony to a pattern of competitive display among villages (Sbonias 1999). Novel motifs are introduced to break the uniformity of seal designs of the previous period, such as new studies of animals and human figures carrying vases (Sbonias 1999, 36; Krzyszkowska 2005: 85). However, it is interesting to note that the preference for schematic decoration, linear or spiraliform motifs is also found on the EM III-MM IA zoomorphic seals whose complex manufacture in three-dimension contrasts with the simplicity of the designs at the base (Galanakis 2005: 24). Tholos tombs A and B at Platanos and the already discussed cemetery at Archanes yielded the most important assemblages (Galanakis 2005: 23). Spirals, leaves and rudimentary representations of animals decorate seals in stone, bone and the so-called “white pieces” (Weingarten 2005: 760; Pini 2000; Karetsou 2000: 302-334), while the use of ivory wanes rapidly. The white pieces comprise a few specimens of scarabs, whose influence is often ascribed to Egypt, and which produced a short-lived stream of imitation on Crete (Krzyszkowska 2005: 74). Interlocking motifs and scenes with animals, insects and a few human figures predominate.

22In terms of style, what is interesting is that a new way of displaying motifs can be observed: an introduction of clock-wise torsions of elements, extended labirinthoid patterns, and line borders along the edges of seals (Wienke 1981). In such manner, principles of syntax are introduced, filling ornaments and reduplication of elements like insects or spirals and parading lions. The tendency to orderly, symmetrical, and schematic arrangement of multiple elements is particularly significant from a cognitive perspective, as it appears to imply a preliminary motion towards narrative schemes (see an example from Platanos, no. 287 in CMS II.1, with the representation of a boat and fish).

23Not only does the scheme of representation change, this phase marks the cusp of development that sees a greatly enriched repertoire of designs, although the attestations for the full MM I period are scanty at best. Despite this paucity, a few aspects can be considered which point towards a progressive ordering of icons on the seal faces that could have constituted a fertile substratum, upon which writing could have formally developed. These aspects begin in the MM I period, and continue progressively and with more prominence, in the succeeding phase. One such aspect is the presence of multiple icons forming a composition that can indicate narrative. Usually these are centred on the human figure in action, engaged in specific activities: men holding jugs, fish, multiple containers, shooting bows, walking, stamping grapes, working at a potter’s wheel (Yule 1980: 119-120). Another expedient is the reduplication of the depicted icons, arranged either in succession on the same sequence, or tête-bêche, facing opposite directions. Items can be reduplicated for emphasis, as seen in the multiple rows of men, vessels, fish, bucrania, spirals, spiders, donkey and goat heads, and cat masks.

Table. Timeline and presence of a sign in hieroglyphic formulas

Table. Timeline and presence of a sign in hieroglyphic formulas

24It is not coincidental that the specific icons that tend to be reduplicated will become (with perhaps the exception of the fish and the spiral) proper script signs in the fully-fledged phase of hieroglyphic use. Of course, the ascription of elements, such as the cat mask or the man, to the proper hieroglyphic repertoire is still a matter of debate. The CHIC opts to exclude the majority (CHIC, 13-14), defining these as only potentially significant, and in various degrees of probability (from non évident to évident, marked with {} or ! in the CHIC, and reported following the same convention in the table). In re-evaluating the CHIC’s choices in selecting signs to be admitted in the standard repertoire, Jasink reintroduces them (Jasink 2009). Part of her reclassification follows Evans, who had originally incorporated many of these in the repertoire (Evans 1909, fig 102-103, such as the full-bodied cat or cat mask, and many others). Whether Jasink’s inclusions are valid across the whole list (several fillers need to be perhaps excluded), it is clear that the correspondences between reduplicated icons and script signs (considered broadly) are not merely casual. Reduplication of icons can work towards a progressive cognitive step from pictographic to ideographic, from an iconic to a symbolic relationship. Reiteration of patterns reinforces the transition, creating a habit in connecting iconic to abstract, between the written sign and the object of the spoken sign (Robertson 2004, 25). In general one can observe in this period a gradual trajectory that prepares the inception of full-blown writing in MM II.

25During this phase, pictorial representations in much more elaborate compositions flourish and the iconographic repertoire becomes rich to the extent that “almost every creature of the land, sea and sky” makes its appearance in the glyptic (Krzyszkowska 2005: 89; see also Yule 1980: pls. 1-18). Full-length human figures, bucrania, quadrupeds, waterbirds, spiders and scorpions populate seal faces on soft and hard sealstones alike. The script becomes fully integrated in this iconographic reservoir.

26The dominant group is the soft-stone group of seals from Mallia. These are more than 700 specimens, the great majority of which are three-sided prisms (Anastasiadou 2013). In her in-depth study of these prisms, Anastasiadou distinguishes the styles of the designs on the basis of three broad types: representational or descriptive (namely figurative images, observable in the physical world), ornamental (abstract and decorative, not observable in the physical world) and hieroglyphic inscriptions (Anastasiadou 2011: 161-23). Some categories can naturally overlap: for instance, floral motifs can take on a decorative as well as a representational character. Anastasiadou also suggests that descriptive images on different sides of one piece could be read together, similarly to images of an ornamental kind (Anastasiadou 2013). There seems to be, in other words, a logical order and a semantic coherence in the way images are arranged on multi-sided prisms.

27Although the use of the script was limited to a small percentage of soft-stone prisms, the narrative schemes found on these is particularly interesting for the prospect of analysing the origin of the Cretan hieroglyphic signs. Statistically, inscriptions on all multi-faced seals interacted throughout their life with iconographical motifs. There is no fool-proof way to tell whether iconography and writing were part of the same narrative scheme, whether the script made sense in conjunction to icons, or whether icons were just decoration, working to further personalise the seals as individual possessions, but not intended to be “read” and understood as part of the specific message to be conveyed. Indeed, the “pictographic” nature of a series of the script signs allows for a polyvalent interpretation: in abstract terms or in absence of a sound methodology, they could be ascribed to the category of script signs and/or representational motifs alike. As Anastasiadou herself claims, “the identification of the nature of the motifs is often made difficult or indeed impossible by the inability to recognise the depicted quantity. This is the result of the lack of knowledge of the cultural and intellectual context which created these devices” (Anastasiadou 2011: 163). Bearing in mind the difficulty in shedding light on grey areas, is there a way to establish a methodology to overcome this inability?

28A strategy to unravel this problem is to study all attestations of isolated symbols which are also part of the hieroglyphic repertoire, as attested on one-sided seals or in isolation on one face of a multi-sided seal: is there a way to garner whether these are to be taken as bone fide inscriptions (albeit formed of one sign only, contravening the standard definition) that are meant to be read phonetically instead of being perceived and understood as simple images?

Phonetic coding and semantic categories

  • 9 In his response to Palaima’s response, Olivier states that ‘any distinction between “true” script (...)

29In establishing a methodology for identifying the core signary of Cretan hieroglyphic, Olivier and Godart start from a typological premise (CHIC, 12-15). The script is not attested only on seals, but it is represented in equal measure on more than a hundred inscriptions on clay supports, such as medallions, bars, crescents and, less frequently, tablets (CHIC, #1-122). Being archival documents, they have no artistic purpose and function only as supports for writing. As a result, they claim that this typology of documents ought to guide the identification of signs, belonging to real writing, to be read phonetically. Seals, on the other hand, may have been engraved with script signs and decoration. As suggested by Olivier, signs on the seals may therefore represent “ornamental writing” (Olivier 1981; 1996; 2000). While at times he brings nuances to his argument, advancing the possibility that the inscriptions on seals are true inscriptions, but that their decorative aspect must always be considered (1990: 14, emphasis Olivier’s), he substantially reinforces the point that seals exist as testament to, or proof of, identification or protection of the objects they identify or protect, rather than carriers of a specific written message. This fails to explain the overlaps in sign-sequences that are present in the archival texts and on the seals, which he (Olivier 1990: 15) and Evans before him (Evans 1909, 260-61), had already noted, but these need to be discussed further because they are less than pervasive. From a terminological perspective, the assumption portended in Olivier’s view, that writing can be ornamental and yet still be writing, seems contradictory: can a decoration be phonologically marked and language notation purely decorative (a point already raised by Palaima in his response, see Palaima 1990: 20-21)?9

30Explaining this apparent paradox is arduous, as one should never lose sight of the fact that the script is undeciphered. However, a thorough analysis of the inscriptions can reveal significant patterns. Models can be helpful in establishing correlations between signs and their possible meaning, especially when faced with undeciphered systems, and these can be functional to assessing whether a proper sign (as opposed to a decorative symbol of iconography) is recorded. The scheme that I have adopted here is based, by and large, on the theory developed by Charles Peirce (Peirce 1931-66), with the specific aim to investigate categories such as pictographic representations, indexical elements, and symbolic or ideographic representations in the corpus of Cretan seals. In order to do this, three sets of data will be analysed:

  • Single sign attestations, engraved in isolation on any one seal face
  • Indexical factors that aid recognition of reading direction or possible indicators of semantic categories
  • Semantic categories that may represent different levels of abstraction, be they defined technically as logograms or determinatives.


31For Pierce, an icon is a sign that stands for its object in virtue of some shared quality, what he calls “likenesses” such as portraits. In terms of writing systems, this relationship is described by the category of the pictogram.

32The majority of the hieroglyphic inscriptions appear on seals and seal impressions of the MM IIA and MM IIB periods. These have been gathered in the corpus of inscriptions (CHIC). Following the principle that an inscription is defined by the close relation of at least two consecutive signs arranged in a coherent sequence and orientation, all single attestations of symbols that have parallels in the script repertoire were de facto excluded from the corpus. Adopting this convention means that a number of single-sign attestations collected in the CMS volumes were intentionally omitted, despite having an identical correspondence to script signs. The authors of the CHIC had taken on the task of producing a list, but left its results unpublished, as they doubted their usefulness (CHIC, 13: n. 27). On the contrary, such an appendix is useful, because the number of single-sign attestations is substantial, if limited in terms of variety. As we built it, we can say that, for the Protopalatial period, the most frequent attestations are of bucrania, men, donkey’s heads, cat masks and boats. All these find correspondences in the conventionalised repertoire on seal faces, although only some turn up in the archival documents. The question remains problematically unanswered: the icon of a man, for instance, may be used as an image, but in another context it may be used as a diagram. If its function is less connected to its being picture-like, and more abstract, how do we differentiate between the two?


33For Pierce, signs whose relation to their objects consists in a relational correspondence or causal connection are termed indices. Most ancient writing systems have a parallel feature in making use of indexical elements that aid reading direction or comprehension. These symbols have been defined as diagrammatical icons (Robertson 2004, 28) on the basis of Pierce’s taxonomy of signs, as they indicate the logic that lies between the interrelated parts represented in writing. For modern writing systems, punctuation marks follow the same indexical, deictic trajectory. Cretan hieroglyphic makes use of two such signs, defined in the CHIC as stiktograms (from the Greek “stiksis”, punctuation, CHIC 13). These are a cross-shaped sign [X] used in word-initial position and a vertical stroke [¨]. Both can be used for separating words, but the former seems to serve different functions depending on the context of attestation. On archival clay documents, it can separate words and also transacted entities, but in certain instances it accompanies the other vertical separator and thus appears redundant (e.g. CHIC #049). Ever since Evans’ day, its function has been attached to reading direction, often marking the beginning of the whole inscription, or emphasising a specific sequence (Evans 1909: 251-253). In this interpretation, it can be seen as “determining the order of many of the groups, notably on the sealstones, where the evidence is not so clear” (Evans 251).

34While this deixis of orientation is conceivably a correct interpretation of the role the stiktogram [X] played in longer sequences, for those composed of several signs forming a spelled-out morpheme, a different explanation ought to be evoked when it appears to sign-post isolated or single hieroglyphic signs. Single signs do not require reading direction to be indicated by punctuation, simply because they are single signs. Yet, on Petshafte and one-face seals such as CMS II.8, 38, CMS VI, 131, CMS VII, 34, CMS X, 280 and others (collected in fig. 6), isolated script signs are signposted by the conventional hieroglyphic stiktogram. The sole extant discussion of this matter concludes that the stiktogram [X] must have functioned as a “motif de remplissage” (CHIC, 13: n. 27). In the same context, it is stressed that the stiktogram may function, exceptionally, to mark a logogram, by preceding it (as is the case for logograms *160 and *176 on a four-sided clay bar, CHIC #53).

Fig. 6. Seals

Fig. 6. Seals

Source: CMS II 8, 38; VI, 131; VII, 34; X, 380 © CMS Heidelberg

35Conversely, it is argued here that this was the case at all times, where this mark was added not as supplementary decoration, but for reasons that are semantically significant. Specifically, stiktogram ¹ could be interpreted as an index for a specific category, most likely as a logogram. The same function could be applied for longer sequences of signs, and it is not a coincidence that this pointer appears when signs on longer inscriptions are isolated or highlighted by decorative fillers, or by dividing lines and bars. When a sign is marked by this pointer, it is to be intended as a logogram. And in this way, even when faced with a single sign on a seal face, the reader is looking not so much at an iconographic emblem to be interpreted, but rather at an actual script sign, as such meant to be read. These cases of isolated signs would be, as mentioned, representative of logograms, but also of monosyllabic words. The distinction can only be revealed by a definitive decipherment.


36For Pierce, signs whose relation to their objects is an imputed character, by virtue of some general or conventional connection between sign and object, are called symbols. This is the most difficult category to investigate when confronted with an undeciphered writing system where distinctions in semantic categories are challenging, if not impossible, to establish. But for instance, patterns of distribution of signs on sealstones can be indicative of some use of logography.

37In the 1960s, Grumach suggested, much along Evans’ line of analysis, that a strong ideographic component was present in the Cretan hieroglyphic script- notwithstanding the existence of phonetic elements (Grumach 1963-4: 375). He sought to justify this conclusion on the basis of affixed and interfixed signs, reduplication of middle signs, and artistic ligatures, which defy a phonetic explanation. To an extent Weingarten shares this view and notices that signs pirouette and alternate by changing their position in ways that do not appear to be “phonetic” (Weingarten 1995). Grumach also introduced the idea that signs appended to attested sequences may function as ideograms or determinatives (Grumach 1968). Signs that are attested in ‘prefixed’ positions, such as sign 042 [ó], sign 038 [f], and sign 001 [~] are indeed likely to represent either ideograms or determinatives, and this inference ought to be supported by a statistical analysis that takes into account characteristics other than the mere sequence-initial position of their attestations.

Fig. 7. Seal

Fig. 7. Seal

Source : CMS XII, 109; VI, 136 © CMS Heidelberg

  • 10 The transcriptions reported in the corpus need to be reviewed. For #255, the arrangement of the si (...)

38Let us consider 042 [ó], the double-axe sign, for instance. This sign appears in sequence-initial position only. The CHIC accounts for three attestations (#255, 113d, #112b) in a medial position, but all of these are doubtful.10 There are other contextual considerations that need to be taken into account that seem to strengthen the hypothesis that this sign works in a different way from the signs that follow it in a sequence. This sign is invariably arranged differently: it is very often rotated 90 degrees from the orientation of the other signs in the sequence, or emphatically singled out from the other signs. The two examples shown in fig. 7 show these sign-posting strategies: in one (a), the double axe appears to be separated, by means of a decorative filler, from the sign that immediately follows it. It is also rotated by 90 degrees in respect to the sequence that follows; this can be interpreted as a logistical ploy to make the wider top section of the signs, where the two blades of the axe extend, fit into the height of the prism. However, this arrangement, also observable on the clay bars and medallions, shows a significant regularity. In the other example (b), the double axe is not rotated, but it is isolated by means of two vertical strokes, virtually encasing it into lines that run along the edge of the prism face. The only instance in which the double axe is neither rotated nor isolated by means of fillers is represented by the Archanes formula, where the double axe is part of a two-sign sequence 042-019, [ó][h].

39Even more substantial is the fact that other signs behave in a very similar way, adhering to particular arrangements that appear to be meaningful, possibly from a semantic point of view. One such example is sign 044 [i], defined by Evans as “perhaps a trowel” (Evans 1909, 187, although he initially identified it with an arbelon: Evans 1894, 36). The “trowel” definition has become standard and been preserved in all literature that adopts iconic terminology to designate Cretan hieroglyphic signs, despite the fact that this sign is less reminiscent of a trowel than, intriguingly, a one-face seal called “Petschaft” (Ferrara & Cristiani 2017).

Deconstructing formulas?

40If arrangements of signs are important in addressing the issue of semantic categories, it is equally important not to lose sight of all the elements present on the seal-face, be they proper signs or additional or supplementary features. Even from a cursory analysis, it is clear that the engraved seal-faces are interspersed with decorative fillers and ornamental devices surrounding the sign sequences. This is an aspect that has been virtually omitted from any discussion of Cretan hieroglyphic, with the exception of recent contributions that highlighted the importance of a holistic approach to the elements engraved on the seals (Jasink 2009; Decorte 2017).

41This calls for an overhaul of the assessment of the sign repertoire as a whole. From a thorough analysis of the material, decorative fillers and signs alike, a number of techniques can be singled out in the layout, composition, display and configuration of all the elements incised on the seal faces. From this analysis and renewed appraisal, one can highlight regular patterns in the organisation of the signs, which can be summarised as follows:

  • Frequent rotation (by 90° or 180°) of specific signs;
  • Emphatic highlighting of specific signs through small decorative fillers;
  • Separation of specific signs via the use of bands or lines;
  • Separation of specific signs via supplementary elements (some of which are proper signs, Jasink 2009) that were deemed “decorative” (CHIC) and excised from the sign repertoire.

42Some of these techniques have already been mentioned above, in relation to sign 042 [ó], but have the potential to be extended to a substantial part of the material, helping thus to reassess the “reading” methodology to be favoured in better understanding the inscriptions, and in providing clearer guidelines to shed light on the “logic” that underpins the composition of the inscriptions. Ultimately, they can be a fundamental tool to distinguish semantic categories, such as logograms or determinatives.

43Indeed, signs that are regularly separated, rotated, or emphasised at the beginning or at the end of sequences could represent determinatives (at the beginning) or logograms (at the end), instead of syllabograms attested within a self-contained sequence. Many decorative fillers such as spirals or scrolls, dots and dividing lines isolate as many signs from the rest of the inscription, assigning to them a particular “reading” emphasis.

44The ambiguity in distinguishing categories does not cease with logograms or determinatives. Jasink suggested that some signs that were omitted in the corpus should be part of the proper syllabographic repertoire (Jasink 2009). Some of these had in origin been included by Evans, and then definitely omitted in the CHIC. These supplemental elements, singled out by Jasink (see Evans’ signs SM 136 –e SM 137, for instance, but many others are pondered over by Jasink) whether or not they are rightfully to be taken as proper signs, seem to occur in specific arrangements throughout the corpus of inscriptions on the seals. These two signs are found within a sequence that is very often repeated, fashioned by the two so-called formulas of signs 044-049 [i][a] and 044-005 [i][C].

45Through a statistical analysis, it is possible to observe that the configuration of these signs is significant when accompanied by these two formulas. Rotation of specific signs is often accompanied by the presence of decorative elements, which aim to encapsulate, in a way not dissimilar from a cartouche, signs from the rest of the sequence. It is a frequent occurrence for a sign to be rotated by 90° while being accompanied by elements that isolate it: in other words, rotation is systematically applied when signs are emphasized, a clue that may strongly hint to the role as logogram played by rotated and filler-isolated signs. Decoration, rotation, and ornamental separation are thus strategies used by the Minoan artists and epigraphers to give emphasis to certain sequences on the sealstones.

  • 11 Note that sign 044 [i] is singled out in the attestations of these sequences even on the archival (...)

46Sign 044 [i] specifically, which is present in the most frequently attested sequences, undergoes all these strategies of singling out, so that it receives emphasis beyond, or aside from, the configuration of the formula. 11 The table brings to the fore this aspect; in almost all sequences of the formulas where sign 044 [i] appears (044- 049 ia and 044-005 [i][C]), this sign is formally separated from the second element, be it by rotation or by “supplementary signs,” or both. Following on from this, that which has been perceived as a unitary formula may need to be in fact deconstructed. The two formulas only very rarely form an inseparable association. For the sequence 044-005 [i][C] it appears that a “simple” configuration, defined as an arrangement whereby the two signs are a) contiguous, b) arranged on the same axis, c) following the same orientation, d) devoid of supplementary signs, e) or of fillers separating them, is absent. Moreover, as customary convention both on the sealstones and on the archival clay documents, the 044 [i] can also be marked by the stiktogram [X], strengthening a logographic reading. All these underlying principles apply to the other sequence 044-049 [i][a], even though less consistently. Statistically, the technical device of “decoupling” is preponderant even in this case.

47These considerations imply that at least the two most frequently found “formulas,” as attested on the sequences on the sealstones, need to be “read” following a methodology that takes into account the technical devices adopted for their overall configurations, which may be significant from a semantic perspective as well as an artistic and cosmetic one. In this light, these stratagems are invested with a deictic role, which guides the reader into addressing and understanding the semantic categories represented. This, ultimately, calls for a novel “reading” methodology that applies these criteria across the whole set of sign-sequences on the seals, especially those of a formulaic nature (Ferrara & Cristiani 2017).

48Half a dozen more “formulas” are attested on the seals only and not on the archival documents on clay, although in a progressively descending order of frequency from the two already mentioned. Statistically, these appear to be arranged more linearly, behaving like proper sequences of signs to be read consecutively. However, this does not mean that they are to be read phonographically. All the formulas found on sealstones, with the exception of the two specimens that show sign 044 [i], are not attested on the archival clay documents, where they could be read as phonographic sequences. Therefore, this could be a tell-tale sign of a possible logographic function, but we cannot prove it incontrovertibly. In any case, this does not detract from the successful endeavour of seeking logographic renderings and postulate that even allegedly contiguous sequences on the seals may have served that very function.

49The scope for logography and the quest to detect it thus brings us back to the evidence from Archanes. The attestations of this formula on the seals found in the necropolis are, as pointed out above, very peculiar. The formula is described as being composed of five signs ([óh/h¢®], Yule 1980; Sakellarakis & Sapouna-Sakellaraki 1997). On the Archanes seals the formula is broken down, always with the first two signs engraved on one face and the rest on the other faces (Karnava 2000: 197; see also fig. 5). This is a regular feature that is attested on several other occurrences. Is there reason to believe that this formula should be deconstructed into at least two different sequences? More importantly, it should be noted that this formula never appears in the archival documents, so far as the extant evidence allows us to see, so the possibility of logography (and not phonographic language notation) should not be discounted altogether. This possibility may be flagged, also in light of the fact that the purported continuity into a Linear A sequence, as contended several times (Olivier 1986; Sbonias 1999; Schoep 1999; Godart 1999; Anastasiadou 2016), is dubious since the signs are not exactly the same. And even if such continuity were to be embraced, albeit with a rough overestimation, it could still be explained as a vestige of a previous period bearing a crystallised significance that transcends the specificities of language notation.

A gradual development and the spark of invention

50The timespan described in this chapter considers a particularly vibrant phase in the development of a fully formed, and ultimately standardised, Cretan writing system, which we can safely ascribe to the script attested on the archival clay documents. This period also saw a boom in the development of seal iconography that showcases a progressively varied reservoir of designs. Within this wide-ranging pool of symbols, a dynamic, if ineffable, interplay marks the transition from picture to sign, and crystallises the birth of writing on the island.

51If it is true that the formulas engraved on the seal faces can, to an extent, be deconstructed, we are facing the challenge of identifying determinatives and logograms as part of the inventory. If in turn it can be proved that a number of signs and sequences as engraved on the seals may be logographic, especially those not attested in the archival documents, then we should expect the reservoir of proper lexemes to be very limited. This, if correct, would not distance the Cretan path to literacy from other inceptions of newly invented writing systems (Mesopotamia, Pittman 1994; Egyptian, Graff & Serrano 2012, etc.). As claimed by Damerow (1999), a writing system does not have to represent the full range of semantics of oral language, especially at its very beginning. In the process of inscribing highly formulaic sequences on seal faces, a crucial transition may thus be witnessed: that of a proto-writing phase turning to full writing, which will eventually culminate in standardised conventions to be set in place, which will then (or even simultaneously) be transposed to the documents on clay.

52How was this turn of events possible, as a process? What we envisage here is a gradual development whose roots are firmly grounded in the Prepalatial period, which generated, at the turn of the MM IA period, an all-new tendency towards an increased sophistication of the iconographic designs. This was mirrored by the ever neater and orderly representations of icons juxtaposed, as part of a new expressive canon, to the earlier geometric and abstract forms, characteristic of the late EM period. Paratactic combinations, repetitions and symmetrical arrangements interacted in a way that is perceived to be highly suggestive of narrative (Anastasiadou 2011). We claim that it was precisely in the dynamic cauldron of logical coherence and orderly iconic configurations that the spark of writing was ignited. This gradual process enabled the natural progression towards conventional codes of language notation.

53In this context, Archanes (if to be read, as we argue here, as highly iconic) stands at the pivot, and embodies the tipping point between proto- and full writing. It would not be surprising if we were encouraged to conclude that the Cretan writing system became fully phonetic only in the transition to the MM II period, a claim that would be strengthened (or rebuked) if we had more evidence from seals in the opaque MM IB period. Despite absences, there is still reason to believe that some form of pictography was part of the iconic literacy of seals for a long time, even beyond the development of the Cretan hieroglyphic system, featuring as a syntax-free, morphologically unmarked element. Be that as it may, we can be confident in stating that the earliest Cretan writing recycled the iconicity of seal designs by harnessing the vibrancy of its contextual expansion and, in this way, triggered a veritable invention.

Appendix. Correlation table signs in parantheses/Cretan symbols

Appendix. Correlation table signs in parantheses/Cretan symbols
Haut de page


The Middle Minoan Three-Sided Soft Stone Prism. A Study of Style and Iconography, Mayence, Philipp von Zabern, coll. « CMS Beiheft 9 ».

–, 2013.
“Of prisms and pictographs: Searching for patterns in MM II glyptic”, Bulletin of the Institute of Classical Studies no 56/1, 129–130.

–, 2016.
“Drawing the line: Seals, script and regionalism in Protopalatial Crete”, American Journal of Archaeology no 120/2, 159–193.

ARUZ JOAN, 2000.
“Artistic change and cultural exchange: The glyptic evidence”, in W. Müller (dir.), Minoisch-mykenische Glyptik: Stil, Ikonographie, Funktion. V. Internationales Siegel-Symposium Marburg, 23-25 September 1999, Berlin, CMS Beiheft 6, 1–13.

“The earliest Egyptian writing: Development, context, purpose”, in Stephen D. Houston (dir.), The First Writing. Script Invention as History and Process, Cambridge, Cambridge University Press, 150–189.

“Il geroglifico cretese: Nuovi metodi di lettura
(con una nuova proposta di interpretazione del segno 044 i)”, Kadmos no 55/1/2, 17–36.

“The origins of writing as a problem of historical epistemology”, Preprint 114 of the Max-Planck-Institut für Wissenschaftsgeschichte, Berlin, disponible en ligne [dernier accès, juillet 2018].

“Cretan hieroglyphic and the nature of script”, in Philippa M. Steele, (dir.) Understanding Relations between Scripts. The Aegean Writing Systems, Oxford, Oxbow Books, 33–56.

“Tomb U-j. A royal burial of Dynasty 0 at Abydos”, in Emily Teeter (dir.), Before the Pyramids. The Origins of Egyptian Civilization, Chicago, Oriental Institute of the University of Chicago, 127–136.

“Primitive pictographs and a Prae-Phoenician script from Crete and the Peloponnese”, Journal of Hellenic Studies no 14, 270–372.

–, 1909.
Scripta Minoa, t. I, Oxford, Clarendon Press.

“The beginnings of writing on Crete: Theory and context”, Annual of the British School at Athens no 210, 27–49.

Minoan Glyptic. Typology, Deposits and Iconography, Oxford, BAR Publishing.

“L’écriture d’Archanès: Hiéroglyphique ou linéaire A ?”, in Philip P. Betancourt, Vassos Karageorghis, Robert Laffineur & W. D. Niemeier (dir.), Meletemata. Studies in Aegean Archaeology Presented to Malcolm H. Wiener, Liège & Austin, Aegaeum, 299–302.

“Segenspfänder”/“Gages de félicité”, in West-östlicher Divan/Divan oriental-occidental, trad. Jacques Porchat, Paris, Hachette.

Préhistoires de l’écriture. Iconographie, pratiques graphiques et émergence de l’écrit dans l’Égypte prédynastique/Prehistories of Writing: Iconography, Graphic Practices and the Forming Process of Writing in Predynastic Egypt, Actes de la table-ronde de décembre 2010, Aix-en-Provence, Préhistoires méditerranéennes, numéro spécial.

“The construction and implementation of the cuneiform writing system”, Visible Language no 15, 345–372.

GRUMACH ERNST, 1963-1964.
“Studies in the structure of some ancient scripts III: The structure of the Cretan Hieroglyphic script”, Bulletin of the John Rylands Library Manchester no 46, 346–384.

–, 1968.
“The Minoan libation formula-again”, Kadmos no 7, 7–26.

HOUSTON STEPHEN D. (dir.), 2004.
The First Writing. Script Invention as History and Process, Cambridge, Cambridge University Press.

Cretan Hieroglyphic Seals. A New Classification of Symbols and Ornamental/Filling Motifs, Pise & Rome, F. Serra.

KARETSOU A. (dir.), 2000.
Kriti-Egyptos. Politismikoi Desmoi Trion Khilietion. Katalogos, Athens, Kapon.

The Cretan Hieroglyphic Script of the Second Millennium B.C. Description, Analysis, Function and Decipherment Perspectives, Thèse de doctorat non publiée, Université libre de Bruxelles.

“Sealstones in cemeteries: A display of social status?”, in K. Branigan (dir.) Cemetery and Society in the Aegean Bronze Age, Sheffield, Sheffield Studies in Aegean Archaeology, 78–86.

Aegean Seals. An Introduction, Londres, Institute of Classical Studies.

Mortuary Behaviour and Social Trajectories in Pre- and Protopalatial Crete, Philadelphia, INSTAP.

Il geroglifico anatolico. Problemi di analisi e prospettive di ricerca, Rome, Dipartimento di studi glottoantropologici, La Sapienza.

“Les sceaux avec des signes hiéroglyphiques. Que lire ? Une question de définition”, in Wolf-Dietrich Niemeier (dir.), Studien zur minoischen und helladischen Glyptik, Berlin, Mann, 105–115.

–, 1986.
“Cretan writing in the second millennium BC”, World Archaeology no 17/3, 377–389.

–, 1989.
“The possible methods in deciphering the pictographic Cretan script”, in Y. Duhoux, T.G. Palaima & J. Bennet (dir.), Problems in Decipherment, Louvain-la-Neuve, Peeters, 39–58.

–, 1990.
“The relationship between inscriptions on hieroglyphic seals and those written on archival documents”, in T. G. Palaima (dir.), Aegean Seals, Sealings and Administration, Liège, Aegaeum, 11–24.

–, 2000.
“Les sceaux avec des inscriptions hiéroglyphiques. Comment comprendre ?”, in Massimo Perna (dir.), Administrative Documents in the Aegean and their Near Eastern Counterparts, Turin, Paravia scriptorium, 141–191.

PALAIMA T. G. (dir.), 1990.
Aegean Seals, Sealings and Administration, Liège, Aegaeum.

Collected Papers of Charles Sanders Peirce, Cambridge, ReInk Books.

Aegyptiaca on the Island of Crete in their Chronological Context. A Critical Review, t. I & II, Vienne, David Brown Book Company.

PINI INGO, 2000.
“Eleven early Cretan scarabs”, in A. Karetsou (dir.), Kriti-Egyptos. Politismikoi Desmoi Trion Khilietion Meletes, Athens, Kapon, 107–113.

“Towards an understanding of the role of glyptic imagery in the administrative systems of proto-literate greater Mesopotamia”, in Piera Ferioli, Enrica Fiandra, Gian Giacomo Fissore & Marcela Frangipane (dir.), Archives before Writing, Rome, Scriptorium, 177–203.

“The possibility and actuality of writing”, in Stephen D. Houston (dir.), The First Writing. Script Invention as History and Process, Cambridge, Cambridge University Press, 16–38.

Archanes. Minoan Crete in New Light, Athens, Ammos.

“Social development, management of production, and symbolic representation in Prepalatial Crete”, in Angelos Chaniotis (dir.), From Minoan Farmers to Roman Traders. Sidelights on the Economy of Ancient Crete, Stuttgart, Franz Steiner, 25–51.

“Looking beyond the First Palaces: Elites and the agency of power in EM III-MM II Crete”, American Journal of Archaeology no 110, 37–64.

“Sealing studies in the Middle Bronze Age, III: The Minoan Hieroglyphic deposits at Mallia and Knossos”, in CMS Beiheft 5. Sceaux minoens et mycéniens. IVe symposium international 10-12 septembre 1992, Clermont-Ferrand, Berlin, Gebr. Mann, 285–311.

–, 2005.
“How many seals make a heap: Seals and interconnections on Prepalatial Crete”, in Robert Laffineur & Emanuele Greco (dir.), Emporia. Aegeans in the Central and Eastern Mediterranean, Liège, Université de Liège, 759–766.

“The invention of writing in Egypt”, in Emily Teeter (dir.), Before the Pyramids: The Origins of Egyptian Civilization, Chicago, Oriental Institute of the University of Chicago, 103–107.

“Typology and style of Prepalatial Seals”, in Ingo Pini (dir.), Studien zur minoischen und helladischen Glyptik, Berlin, Mann, coll. « CMS Beiheft 1 », p. 251-261.

YULE PAUL, 1980.
Early Cretan Seals. A Study of Chronology, Mayence, Philipp von Zabern.

Haut de page


1 I am very grateful to John Bennet for having commented on this paper, and for his constant and unfailing support over the years. Gratitude also goes to Miguel Valério for his many suggestions and inspirational insight. All shortcomings and inaccuracies are my own. This chapter follows the conventional period division established by Arthur Evans for Minoan chronology: EM= Early Minoan (corresponding very roughly to 3100-2000 BCE), MM= Middle Minoan (2000-1700 BCE), LM= Late Minoan (1700-1100 BCE).

2 Olivier and Godart (1996) point out that the specimens from Samothrace, Knossos and Gouves only bear the Archanes formula, but are not likely to belong to the “Archanes script”, cf. CHIC (Corpus Hieroglyphicarum Inscriptionum Cretae, Olivier & Godart 1996) 18: note 59.

3 Recueil des inscriptions en linéaire A, Godart & Olivier 1976-1985.

4 The Cretans may well have had awareness of Egyptian literacy, but we are in the dark as to whether their perception matched their understanding of it. Exchanges in material culture were already underway. For seals, they date as early as the Middle Kingdom, when the earliest Egyptian scarabs and scaraboids appear on the island (Pini 2000; Weingarten 2005). The earliest seals from this period are mostly made of hippopotamus tusk, whose source was probably Egypt. This import was so significant that it contributed to a boost in the craft of seal engraving in EM III-MM IA (Krzyszkowska 2005). The Cretans soon started to produce imitations and adaptations of the Egyptian models (Aruz 2000), cultivating a vibrant, if short-lived fashion in response to the exotic imports (Phillips 2008). The Mesara region is predominantly involved in this manufacture, especially of the so-called “white pieces,” which imitated the appearance of ivory with a coating of magnesium silicate (Pini 2000).

5 Cat. no. 50 in Phillips 2008: HM 464. Cf. also CSM II.1: 468, no. 395.

6 In the same building, a bone seal, in the shape of a fly, seems to be slightly earlier (EM III-MM IA) than the scarabs, of Minoan manufacture rather than an import, and probably locally inspired, although the symbol of the fly was common in Egypt. Particularly intriguing is the engraving on the base, which is tantalisingly reminiscent of motifs that may be reconnected to the Cretan hieroglyphic script (fig. 4). To this end, it must be noted that this is not taken to be an inscription in the corpus, but it may represent “pseudo-signs” at a prototypical stage. In fig. 4, some tentative correspondences with known signs from the slightly later Cretan hieroglyphic repertoire are drawn, but such suggestion has no affirmative value, and as such remains entirely speculative.

7 Phillips maintains that the typology of sistrum found at Archanes has a closed, arched top found in Burial Building 9 and represents a style that does not enter the Egyptian material culture before the New Kingdom. This makes it unlikely that the prototype for this object is Egyptian, and that the influence may have been ‘in the reverse direction’ (Phillips 2008: 36).

8 We would like to warn the reader that every letter in blog and in parantheses correspond to a sign which is available in the table at the end of the article.

9 In his response to Palaima’s response, Olivier states that ‘any distinction between “true” script and “decorative” script may be misleading. Sumerian and Egyptian are true scripts from their very beginning, but they have full “pictorial” or “decorative’ characteristics” (Olivier in Palaima 1990, 22), which seems to imply that Cretan hieroglyphs are only formally pictographic, but this is by no means made clear.

10 The transcriptions reported in the corpus need to be reviewed. For #255, the arrangement of the signs is completely arbitrary and sign 042 happens to be placed close to the rim of the seal, suggesting that it is indeed in sequence-initial position. The corpus records the whole set of signs, six in total, as one sequence, which is equally doubtful.

11 Note that sign 044 [i] is singled out in the attestations of these sequences even on the archival documents. Even on these configurations, which represent more schematic, linear attestations, the scribes often made a point of ornately dotting the sides of the sign for emphasis.

Haut de page

Table des illustrations

Titre Fig. 1. Phourni Cemetery Site Map, Archanes, Crete
Crédits Croquis : Silvia Ferrara
Fichier image/jpeg, 197k
Titre Fig. 5. Multifaces seals of Archanes
Fichier image/jpeg, 181k
Titre Table. Timeline and presence of a sign in hieroglyphic formulas
Fichier image/jpeg, 508k
Titre Fig. 6. Seals
Crédits Source: CMS II 8, 38; VI, 131; VII, 34; X, 380 © CMS Heidelberg
Fichier image/jpeg, 144k
Titre Fig. 7. Seal
Crédits Source : CMS XII, 109; VI, 136 © CMS Heidelberg
Fichier image/jpeg, 140k
Titre Appendix. Correlation table signs in parantheses/Cretan symbols
Fichier image/jpeg, 124k
Haut de page

Pour citer cet article

Référence électronique

Silvia Ferrara, « From icon to sign », Terrain [En ligne], 70 | octobre 2018, mis en ligne le 06 novembre 2018, consulté le 16 novembre 2018. URL : ; DOI : 10.4000/terrain.17225

Haut de page


Silvia Ferrara

Université de Bologne

Haut de page
  • Logo CNRS
  • Logo Université Paris Lumière
  • Logo Ministère de la Culture
  • Logo Université Paris Ouest
  • Logo MAE - Maison Archéologie Ethnologie
  • Logo FMSH-Diffusion
  • OpenEdition Journals