Of Tables, Rakes, and Blankets in the Indus Script

Seal M-153 with inscription: DUCK HEAD / MAN HOLDING COMB / BIRD BETWEEN PARENTHESES.

Continuing with my enumeration of the Indus signs, I term the eleventh among those drawn with eleven strokes MAN HOLDING COMB (XI 11).  The object in the “man’s” hand could just as easily be called a “rake,” as there is a line from the end of the “arm” to the “comb.”  This attaching stroke is diagonal as it is in most other cases (e.g., MAN HOLDING CUP; MAN HOLDING FOOT, etc.).  In the case of each of these other symbols, this connection is not part of the object, as demonstrated by the existence of the independent signs CUP, FOOT, and so on, without an attached stroke.  Thus, I assume the object held in XI 11 is the COMB – which lacks such a connecting stroke -- rather than the RAKE, which requires a “handle.”

Bangle M-1639 with inscription: BOAT / MAN HOLDING 4 QUOTES (?) / STRIPED BISECTED TOP.

In any case, this symbol appears in other lists as KP26, W15, and Fs A-22.  Fairservis also sees this as a depiction of a man with a comb, defining it as “scribe” for reasons having to do with near homophones in Dravidian (“comb” being cippu, somehow associated with Kannada kiri “to shave,” which sounds like kiru “to scratch, mark, write,” hence the connection to a writer or scribe) (1992: 35).  Wells gives the total occurrences of this sign as seven, all but one from Mohenjo daro.  One instance listed in Wells’ catalog is probably an error: the sign on the seal from Lothal is MAN HOLDING QUOTE, with an unattached BI-QUOTES on the other side of the “head” of the anthropomorph (L-12).  A second instance may also be erroneous, though I am less certain of it.  Occurring on a bangle, it is marred by a crack passing through it (M-1639).  In this case, I see four short strokes above the connecting stroke that rises from the arm, but no horizontal line beneath the four “quotes.”  Thus, I take it to be an instance of MAN HOLDING FOUR QUOTES.

Detail of inscription from broken seal L-12: FIVE POSTS / MAN HOLDING QUOTE /
BI-QUOTES // CORN HOLDER (WITH ATTACHED LOOP?) / CRAB.

Wells subdivides this sign into three variants based on the number of “teeth” in the “comb.”  His “a” has five “teeth,” the central one appearing to be an extension of the connecting stroke; his “b” has four “teeth”; and his “c” has only three.  The connection between the “comb” element and the arm is not diagonal but vertical in the case of “c,” for which reason I took this to be an anthropomorph with the RAKE.  But Wells may be correct in grouping all of these together.  (Fairservis appears to view Wells’ “c” as “a man holding a stalk of wheat,” in my terms holding a TRI-FORK (A-26; 1992: 35).

Rock art motif of a man with rayed arc overhead from British Columbia (Keyser 1992: 51, fig. 24a).

The closest parallel to this Indus sign that I have observed comes from the rock art of North America.  Stick figure humans are common in many places, as is an element much like the Indus COMB (though termed a “rake” in the literature).  However, these two elements seldom join in a way that would appear to show the anthropomorph holding the “comb.”  In the Columbia Plateau region (the area of Washington, Idaho and British Columbia around the Columbia and Snake Rivers), there is a common motif combining an anthropomorph with a “rayed arc” (Keyser 1992: 61-62, figs. 35 and 36).  In most cases, the arc is either attached to the figure’s head or is just above it.  But occasionally, the arms of the anthropomorph bend at the elbow and the arc nearly descends to the ends of these arms, as if the human were holding it (e.g., fig. 35b). 

Rock art depiction of a supernatural being with comb-like hands in the Long Narrows style (Keyser 1992: 89, fig. 61c).

Depictions of supernatural beings are frequently found in the Long Narrows style of the Lower Columbia region.  A few of these show a comb-like element at the end of each arm (e.g., 1992: 88-89, figs. 60d and 61c).  In these cases, the “comb” most likely represents a hand.  In both the examples cited, each comb/hand has exactly five teeth/fingers.  This suggests one possible interpretation of the Indus signs of RAKE and COMB.  The former may depict a highly schematic arm, while the latter could represent just a hand.  These seem rather unlikely explanations, however, since both Indus signs frequently have fewer than five prongs (specifically, three or four) and occasionally more (six or seven).  Still, such inaccurate renderings of hands or feet, with the wrong number of digits, do occur in less stylized traditions.  Human footprints in Mixtec manuscripts sometimes have such inaccuracies (e.g., Smith 1973: 225, fig. 13a which includes a series of footprints showing three, four, and five toes).

Broken seal Pbm-1 with partial inscription: TWO POSTS (?) / POTTED ONE / RAKE / FISH / BI-RAKE / FISH (??).

The second Indus sign discussed in this post is BI-RAKE (XI 12), a symbol also enumerated KP95, W266, and Fs Q-12.  In appearance, it resembles the basic RAKE, but there are two rows of “tines,” one row placed over the other.  The central “handle” connects both these rows and descends beneath them.  Obviously, if this is related to the basic RAKE, it is hardly likely to represent an arm, since two hands do not appear on a single arm.

Cylinder seal from Susa, Iran with partially legible inscription including BI-RAKE / BIG SHOULDERED MAN / STACKED 12 (?) / DOUBLED BELTED DOUBLE AITCH / (perhaps a FISH also; over repeated bovines at the bottom). 

Fairservis considers this symbol to be essentially the same as the basic RAKE, doubled stylistically, and takes it to be a proper name.  But in at least one inscription, both the basic RAKE and the "doubled" BI-RAKE both appear, a fact which suggests they are independent symbols (see Pbm-1).  Wells, avoiding definitions, finds 15 occurrences, with 10 of them from Mohenjo daro (though I find 13), three from Harappa (I see 7), one from Lothal (I see 2), and one from Pabumath.  A BI-RAKE also appears in the Harappan-style inscription on a seal from Susa, Iran (Collon 2005: 143, fig. 608).  Wells and I also differ in the number of variants of this symbol.  He sees two, with the only difference being the number of “tines” on the uppermost “rake”: his “a” has five, while his “b” has four.  I would add a “c” with eight (M-741) and “d” with only three (H-701).  There might even be an “e” since the example from Susa seems to have six.

Proto-cuneiform sign NUN~a, "prince," which also occurs doubled and stacked, becoming NIR~a, also "prince."

In proto-cuneiform, there is a sign that is more like the Indus SHISH KEBAB, with the designation NUN~a.  This symbol also occurs doubled, then being transcribed NIR~a.  Both signs came to mean “prince.”  If this principle of doubling without a change in meaning applies to Indus signs (and there is no way of knowing whether it does), then Fairservis may be correct in seeing the BI-RAKE as nothing more than a fancier version of the RAKE.  I am hesitant to accept this, however, since a clearer example of a doubled RAKE is RAKE OVER RAKE (a 12-stroke sign to be discussed later).

Bar seal M-1271 with inscription: LOOP ARMED MAN HOLDING SLASH / CEE BOAT /
STRIPED FAT CHEVRON / STRIPED VEST / STRIPED TRIANGLE / SPACESHIP UNDER TABLE / COMB.

The following sign is SPACESHIP UNDER TABLE (XI 13), also known as KP222, W422, and perhaps to be viewed as Fs N-1, “mountains” under I-19, “sky; superior.”  However Fairservis saw it, there are three instances, as Wells notes, all from Mohenjo daro.

Proto-cuneiform |SILA3~a x KUR~a|, a combination of "market" and "(foreign) land."

The three stacked wedges denoting “mountain, (foreign) land” in proto-cuneiform are cited by Fairservis in his definition of the SPACESHIP element.  Thus, the proto-cuneiform |SILA3~a x KUR~a| may be considered more or less analogous to SPACESHIP UNDER TABLE.  Instead of a “table,” the proto-cuneiform includes a “greater than” shape with a wedge at the bend.  This portion came to mean “market, square.”  This makes one wonder whether the combination represented a foreign market.

Detail from seal M-896 with inscription: STRIPED TRIANGLE UNDER DOUBLE TABLES /
SINGLE QUOTE // TWO POSTS / MAN HOLDING POST / BEARER.

Our next sign is STRIPED TRIANGLE UNDER DOUBLE TABLES (XI 14).  It appears in the literature as KP213 (with 12 strokes) and W429 (also 12 strokes).  The actual sign is a bit difficult to make out in the photograph in the Corpus, but I think it may have only 11 strokes.  In any case, it is a singleton from Mohenjo daro (M-896).  It is also the only instance with a TABLE over another TABLE (over a third element).

Proto-cuneiform sign |SILA3~c x ZATU 687|, a combination of "market" and an unknown sign.

If the SILA3 cited under the previous sign may be taken as equivalent to the Indus TABLE, then there is an analogous sign in proto-cuneiform.  It includes a triangular element with a rounded base, rotated 90 degrees, beside the “greater than” element.  This combination is transcribed |SILA3~c x ZATU 687|.  Similar rounded triangles represent oil or a dairy product such as butter.  But this one, with both a vertical stripe and a horizontal stripe inside, may be distinct.

Seal M-1169 with inscription: TWO POSTS / BLANKET WITH 7 TICKS / BIRD BETWEEN PARENTHESES / BI-QUOTES // SKEWERED CHEVRON / FISH // (2nd row) TWO POSTS / FISH / CUPPED POST / THREE POSTS / SPEAR.

The fifth Indus sign is BLANKET WITH SEVEN TICKS (XI 15).  Only Wells notes this as an independent sign (W538), another singleton from Mohenjo daro (M-1169).  Fairservis considers all the BLANKET variations to represent enclosures, suggesting further that the varying marks inside indicate varying weights (1992: 100).  In his discussion of these variations, he shows BLANKETS with four, six, and eight “ticks” first, but also includes one with asymmetrically arranged seven (four at the top, three at the bottom, the same arrangement of strokes as in most occurrences of STACKED SEVEN).

There are certainly many square or rectangular symbols outside the Indus Valley.  But while it is easy enough to find grids, or internal stripes, or even quadrilaterals enclosing a cross or “X,” I have seen none with “ticks” elsewhere.  This symbol may be a unique type, found only in the Indus script.  It may be of interest to note, in addition, that while grids or checkerboards are among the entoptic shapes seen by those in altered states of consciousness such as trance, an element such as the BLANKET is not.

Broken seal M-637 with (partial?) inscription: BLANKET WITH 4 TICKS, 3 HYPHENS / POT.

The next sign may be a variant of the last: BLANKET WITH FOUR TICKS, THREE HYPHENS (XI 16).  Once again, only Wells enumerates this separately (W534), finding it to be yet another singleton from Mohenjo daro.  Fairservis does note its existence in his full discussion of the BLANKETS.  It could be that the internal marks represent numerals, as Fairservis suggests.  But if that were the case, one would expect instances with “one,” “two,” and “three” internal strokes.  As it is, “four” is the lowest apparent numeral and “ten” the highest.  In addition, some BLANKETS also include elements that differ from the usual “quote.”

Detail from seal H-598A with inscription: DUCK HEAD IN LEAF TOPPED POT / PRAWN /
ZEE / BI-QUOTES // WHISKERED FISH / BLANKET WITH 4 TICKS, 2 HYPHENS, & CEE / POT.

There is one more similar sign: BLANKET WITH FOUR TICKS, TWO HYPHENS, AND CEE (XI 17).  It appears in other lists as KP270(c), W536 and 537, and Fs G-15.  Despite his suggestion that the internal marks are numerical, Fairservis defines this symbol as representing a proper name.  Wells, for his part, gives two versions: W536 containing the 4 “ticks,” a backward “C” shape, and 2 “hyphens” of the name, while W537 contains the same “ticks” and “cee” but then two “quotes.”  Since both are singletons from Harappa, I have combined them (somewhat arbitrarily) as two variants of a single sign.

Detail from seal C-8 with inscription: STRIPED BATTERY WITH ATTACHED LOOP / POTTED ONE / BELTED FISH / CIRCLED TRI-FORK / POT (note that the small LOOP is not an independent sign, but part of a ligature here).

After these BLANKETS, I include an eleven-stroke STRIPED BATTERY (XI 18).  It appears only in Wells’ list as an independent form (W477”c”).  Both Fairservis and the team of Koskenniemi and Parpola recognize a striped version of the BATTERY, but in both cases the stripes are mostly vertical (KP290, Fs G-2, my X 10).  The version with eleven strokes has only horizontal stripes.  As such, it seems to occur twice, once at Mohenjo daro and once at Chanhujo daro.  However, the photographs clearly show that in both of these, there is a loop attached to the BATTERY which Wells failed to include.

Three variants of proto-cuneiform URUDU, "copper," analogous to the Indus (STRIPED) BATTERY.

Proto-cuneiform provides good analogies to the STRIPED BATTERY portion – though without any attachment.  The sign URUDU includes a variant with some vertical and horizontal stripes (variant “d”) and one with diagonal stripes (variant “c”).  These came to mean “copper.”

Seal L-86 with inscription: GRID (2 X 4) WITH ATTACHED POST / BACKSLASHES IN OVERLAPPING
CIRCLES / TRI-FORK (note the rough appearance of the signs, with the GRID leaning to one side).

Another sort of attachment appears in the following Indus sign: GRID WITH ATTACHED POST (XI 19), also known as KP269 and W505.  It is interesting to note that the sign presents a very neat appearance in both of these published lists, with perfectly vertical sides.  On the actual seal from Lothal, the GRID element has a definite tilt to one side, though.  It should actually be placed among the signs with ten strokes (the GRID including 2 x 4 internal squares), but I include it here due to its appearance in the list of Koskenniemi and Parpola (including 3 x 4 internal squares).  Grids are common around the world, perhaps due to their inclusion among the entoptic shapes, but with such an attachment this sign becomes unique.

Inscription from bar seal M-391: FOUR TOED FOOT / STACKED STOOLS WITH MID POST / VEE IN DIAMOND / BI-QUOTES // CIRCLED E TRI-FORK / BACKSLASH IN FISH / 3 QUOTES / CIRCLE BETWEEN CEES / POT // COMB BELTED ASTERISK / SLASHES IN OVERLAPPING CIRCLES / QUAD-FORK (two units of info may be included).

Today’s final sign is COMB BELTED ASTERISK (XI 20), appearing elsewhere as KP249 and W435.  It is a singleton from Mohenjo daro (M-391), combining the “X” with doubled “quotes” in the upper and lower parts (the ASTERISK of eight strokes), with a “comb” across the middle that includes four “teeth.”  As with some of the other ligatures discussed in this post, the individual elements have analogs in many other scripts and art traditions, but I have seen no such combination elsewhere.

Indus Signs of Eight Strokes: The First Eight

Tablet H-817 with inscription (right to left): STACKED EIGHT A (4 X 4) /
STRIPED VEST / POT / COMB (note the neat stacking on this bas-relief tablet).

Of signs comprising eight strokes, the first is the simplest, STACKED EIGHT (VIII 1).  It contains two rows of four short vertical lines each, one row over the other.  In the lists of other scholars, this rare symbol is identified as KP128, W208, and Fs O-12.  Fairservis sees in this sign the numeral eight, specified as being an adjective (which means it modifies a noun).  Apparently because of similar-sounding words in Dravidian languages, he further defines the sign as “count, number; pair, couple.”  He cites seven occurrences (1992: 62).  Wells, on the other hand, finds only six, three from Mohenjo daro and three from Harappa.  I see seven, adding one from Lothal to Wells’ total (L-125).

Detail from iimpression of seal M-83 with inscription (right to left):
CIRCLE WITH THREE LASHES UNDER TABLE / STACKED EIGHT A / TRI-FORK
(note the grid-like element beneath the head of the "unicorn," the top of the cult stand).

Proto-cuneiform includes at least eleven different ways of writing “eight,” all of which are identified with the letter “N” (for numeral) plus a number (N01, N02, N08, N14, N18, N19, N20, N34, N35, N51, N57, N58).In this proto-writing system, the different types of numerals enumerate different types of things.  The set for grain differs from that for herd animals, for example.  In almost all types, there is a column of four identical marks beside another such column.  Only N57 differs from this.

Three types of "stacked 8" numerals from proto-cuneiform: 8(N14) at top, 8(N01) in center, and 8(N18) below
(all should be rotated 90 degrees, forming two columns of four identical marks each).

Proto-Elamite has fewer distinctions in its known numeral systems.  But it too enumerates things differently based upon the type such as grain versus animals.  According to some experts, these distinctions in writing mirrored similarly distinctive ways of counting things (Schmandt-Besserat 1992:185-188).  Some of these ancient counting systems were base ten, like the modern English numerals, while others were not.  The larger numerals were often base sixty, for example, while some of the smaller sets were base six.  Fairservis thought that the Indus numerals might have been base eight, because there are many instances of STACKED SEVEN but few occurrences of larger numerals.  In a previous post, I pointed out some of the problems with this view, in particular the large group of STACKED TWELVE occurrences (70 or so).

Seal H-65 with inscription: STACKED EIGHT B (3 X 3 X 2) / DOUBLE BACKSLASHES / POT.

The next Indus sign may or may not be a variant of VIII 1.  Since there is disagreement among the experts, I tentatively call this one STACKED EIGHT B and assign it a separate numerical designation, VIII 2.  It is found elsewhere as KP141(a), W229, and Fs N-2(b).  Wells shows three variants (though he states that there are only two occurrences -- a misprint?).  His variant “a” is made up of three stacked rows of dots or short verticals (3 x 3 x 2).  His “b” is similarly arranged, but not as neatly (3 x 3 x 2 again) and “c” is inverted (2 x 3 x 3).  His “a” is a singleton from Harappa (H-65), his “b” a singleton from Mohenjo daro (M-1078).  The inverted variant “c” appears twice at Harappa (H-322 and H-967).  By my count, this adds up to four occurrences, not two as he first states. 

Tablet H-322 with inscription (right to left): STACKED EIGHT (2 X 3 X 3) /
TWO POSTS / POT / COMB (initial sign VIII 2 is variant "c").

This is not a criticism of Wells, but an observation.  It is not always a simple matter to decide just how many there are of a given sign, especially when dealing with apparent numerals.  Sometimes the apparently non-numerical BI-QUOTES (resembling double quotation marks) appear above another sign, due to crowding.  When the BI-QUOTES occur above a number of hash marks, one must decide whether to count the two strokes of the BI-QUOTES in with the strokes beneath -- or not.  I am guessing that Wells did not originally count the inverted STACKED EIGHT as a version of "eight."  Instead, he probably separated out the two strokes at the top as an instance of BI-QUOTES, counting the lower strokes as STACKED SIX.

Tablet H-967A with inscription (right to left): STACKED EIGHT B (2 X 3 X 3) /
DOUBLE POSTS / POT / COMB (note how untidy the "eight" is here compared to the seal above). 
Here, one might consider the eight strokes on the right to be BI-QUOTES (over) / STACKED SIX,
except that the BI-QUOTES almost never occurs in initial position anywhere else.

Fairservis, on the other hand, only notes one variant (3 x 3 x 2 like Wells’ “a” and “b”), but the strokes in it are tilted.  It is, then, three slashes over three backslashes over two slashes.  He does not consider this a numeral at all, but a depiction of a river or stream, meaning “water.”  It does vaguely resemble the Old Chinese shui3, “water,” which may depict a curving rivulet plus four smaller whirls of water (Wieger 1965: 287).  The Indus STACKED EIGHT (both VIII 1 and VIII 2) also recall one form of the Old Chinese yu3, “rain” (a stacked “six” over four “posts,” all under a “roof”) (1965: 288).  Fairservis even refers to the latter when he makes his case for interpreting the STACKED TWELVE as rain, a point we will get to in another post.

Old Chinese "water" (left) and "rain" (right).

Before we leave these apparent numerals, it is worth noting that there are no occurrences of *EIGHT QUOTES or *EIGHT POSTS (the asterisk indicates that it is not attested), although every smaller number appears to be depicted by one or both method, as does “nine.”  Thus, the Harappans do not seem to have cared much for “eight.”  In contrast, we previously noted four occurrences of SEVEN QUOTES, two of SEVEN POSTS (though both somewhat questionable), plus 38 instances of STACKED SEVEN, adding up to 44 “sevens.”  This difference -- 44 "sevens" versus only of a type often in folklore (as noted in the post on apparent numerals).

Detail from abraded seal M-966 with inscription: BOAT / PINCH / POT (?) / WATERY SEVEN (?) /
POT (?).  Several signs may be read in more than one way, due to the poor condition of the seal. 
Thus, it might rather be CEE / PINCH / POTTED ONE / STACKED THREE /
ESS (SINGLE POST?) / STACKED FOUR / CUP.

The third Indus sign looks more like Chinese “water” than the apparent eight (or the "twelve" for that matter), so I tentatively call it WATERY SEVEN (VIII 3).  In form, it is a STACKED FOUR on one side and STACKED THREE on the other side, with the two "numerals" separated by a long stroke.  In KP155, the central divider is shown with a slight “S” curve, whereas in W224 this separating element is straight.  An examination of the photo of this sign in the Corpus is not particularly revealing when it comes to deciding the exact form of the long stroke, because the seal on which it occurs is in such poor condition.  Still, it seems to me that this group of marks might be read as more than one symbol with equal justification, say, as STACKED THREE / ESS / STACKED FOUR. 

No one refers to the enigmatic WATERY SEVEN as a numeral, to my knowledge.  However, if one considers POSTS and QUOTES to be numerical symbols, one might reasonably conclude that this is a group of three numerals, FOUR + ONE + THREE and thus “eight” -- so perhaps my term for it ought to be WATERY EIGHT.  In a way, though, it does not really matter what one calls it, because there is only this one.

A few additional remarks concerning numerals are in order here.  Some amateur enthusiasts interpret the many instances of CUP + POSTS/QUOTES as numerals, as if they were Roman numerals (VII, VIII, and so on).  But this type of stroke combination does not behave quite like the superficially similar Roman numerals.  For example, there are instances of VI (CUP + SINGLE POST), VII (CUP + DOUBLE POSTS), and VIII (CUP + THREE POSTS).  Indeed, these resemble the Roman six, seven, and eight.  But among the Indus inscriptions, there are also several occurrences of VIIII (CUP + FOUR QUOTES/POSTS), while the Romans switched to writing IX for "nine."  But there are other inscriptions that make things even worse for the hypothesis that "V" = 5.  There are some combinations of VIIIII (CUP + FIVE QUOTES/POSTS), and even a few VIIIIII (CUP + SIX POSTS/QUOTES).  The Romans might have written "V" plus four posts once upon a time, but they never wrote "ten" as "V" plus five.  They had "X" for that, so that was likewise no "VV" (which, by the way, does occur among the Indus inscriptions).  One might then suggest that the Indus "V" was not "five" but "ten" (or even "100").  But there is no instance of a CUP (whether “U” or “V”) plus eight strokes, whether “quotes” or “posts.”  Why would the Harappans count up as many as 16 items but never mention 18 or 19?  It is too difficult to explain such oddities, so I conclude that the CUP (“V” or “U”) is not a representation of “five,” or any other numeral, despite the superficial resemblance to a Roman numeral.

Proto-cuneiform GI6, "to be black, dark" (resembling a "stacked 8" under a chevron).

As a comparison, there is a sign in proto-cuneiform that seems to incorporate a numeral, at first glance.  The sign GI6, “to be black, dark,” is a “stacked eight” beside (or under) a “less than” sign (< + four stacked ­ + four stacked ­).  This symbol is similar to one version of Old Chinese “rain,” which also incorporates four or ten short strokes that might be interpreted as numerals by a naive viewer.  In other words, while something may appear numerical to us, that does not prove it is.

Detail of seal K-13 with inscription: BLANKET WITH FOUR TICKS / POTTED ONE /
STACKED SEVEN / EF PRONGED EXIT / POT (over unicorn and stand).

The next symbol is one of many that I term a BLANKET.  It is a simple rectangle, adorned with two short marks protruding down from the top horizontal and two extending upward from the base.  To distinguish this version with four inner strokes from instances with other numbers of additions, I call it BLANKET WITH FOUR TICKS (VIII 4).  It is also KP270(b), W529, and Fs G-11.  Fairservis considers the marks enclosed by the rectangle to be significant.  The whole symbol means “enclosure,” he suggests, with the “ticks” diacritical marks to be seen as representing suffixes – at least when found in combination with the ubiquitous POT symbol.

Looking to Old Chinese once again, the character for “water” sometimes occurs inside an oval or rounded rectangle.  Regardless of the shape of the surrounding symbol, the resulting character does not mean “enclosure.”  Instead, it is yuan1, “a whirlpool, a gulf, an abyss” (Wieger 1965: 288).

In support of Fairservis’ interpretation, though, we might point out the Luwian hieroglyph DOMUS, “house.”  This is a square or rectangle with two additions inside, attached to either side.  The “diacritical” marks in this case are a “C” shape on the right and the reverse on the left.  In other words, these are not diacritical marks at all (not signaling the presence of suffixes).

Luwian DOMUS, "house."

The Egyptian glyph of a wall is also reminiscent of Indus VIII 4 (O-36).  It is usually a rectangle that is taller than it is wide (though occasionally wider than tall), with added, thicker strokes spaced around the perimeter.  It is an ideograph or determinative in the word inb, “wall,” in sbty, “surrounding wall,” and in wmtt “bulwark, fortification.” 

Analogs to Indus VIII 4: proto-cuneiform |GA2~a x 3 (N57)| (upper left); Egyptian glyph O36, a determinative in wmtt, "bulwark, fortification" (center top); Cretan "gate" hieroglyph (upper right); Old Chinese yuan1, "whirlpool" (lower left); proto-Elamite M145~c (center bottom); and Cretan TELA, "cloth" (lower right).

The Cretans also used hieroglyphs, early on, and a rectangle with about four “ticks” protruding downward from the base is the ideograph TELA, “cloth.”  The rectangle also contains a backslash running from the upper left corner to the lower right.  Besides its ideographical usage, it may also signal the syllable wa.

Proto-cuneiform includes a sign made up of a rectangle wider than it is tall with a stacked set of three horizontal strokes and a single crossing vertical.  This is another of the signs that seems to contain a numeral and it is transcribed as if that were the case: |GA2~a x 3 (N57)|.  However, it came to have the non-numerical meanings “box, basket; house; stable; shrine.”  Proto-Elamite also has symbols that are rectangular and contain inner strokes (M147 with four short horizontals and two central verticals, M145~c divided into quarters like a window and with two additional strokes in the upper right and lower left sections, and M145~e which is the reverse of the previous).

Adinkra cloth showing grid-like patterns, nkyimu above and kronti ne akwamu below.

Square or rectangular symbols also occur in Africa, where they are not part of a writing system.  The Adinkra symbol nkyimu is one of these, subdivided into several smaller squares or diamonds.  The name refers to the divisions of the Adinkra cloth before it is printed, “blocking.”  The symbol represents adroitness and precision (Willis 1998: 146).  Here, the precise configuration and number of internal strokes can vary quite a bit without changing the meaning of the overall symbol.  This may also have been true of the Indus BLANKET, which seems to have multiple variations with varying numbers of TICKS.

Seal M-74 with inscription: LOOP ARMED MAN HOLDING SLASH /
CUPPED SPOON / BATTERY / TRIPLE STRIPED MALLET.

The next sign in my list is TRIPLE STRIPED MALLET (VIII 5), also known as KP280 and W470(b).  We have seen “mallets” before, without any internal marks and with just one or two.  The form with three stripes occurs at least twice (M-74 and M-1203).  It may be compared with the mortar and pestle used by a man in Egyptian glyph A34, the wagon or winnowing shovel of proto-cuneiform MAR~a@t (though this has seven stripes which do not cross the whole thing), and proto-Elamite M167~a.  One might also consider the Egyptian glyph of a column with a tenon at the top to be a reasonable parallel (O28).  Less similar is Luwian glyph SCRIBA, “scribe.”  This might be described as an ear of grain under a "table," or a peculiar, inverted "mallet."

Inscription from Lh-1: MAN HOLDING DEE-SLASH / MALLET / BI-FORK / BI-QUOTES //
FISH UNDER CHEVRON / FOUR-TOED FOOT / MAN BY CHEVRON / GRID (2 X 4?) /
BI-FORK TOPPED POT / POT (the original is less than perfectly clear).

Sign VIII 6 is another GRID (sometimes referred to as an ICE CUBE TRAY in my earlier lists because it has two columns of stacked squares).  I specify the number of internal divisions with numbers, 2 x 4, to indicate two columns, each of four spaces.  It is elsewhere KP268(c), W502, and Fs G-17.  Fairservis suggests it represents another enclosure, this one divided into compartments, and probably a variant of his G-16, “house.”  I designate his G-16 as a different GRID (3 x 4).

An unbounded grid as seen in American rock art (left) and proto-cuneiform E2~a, "house" (left).

Egyptian hieroglyphs include an unbounded, rectangular grid, representing land marked out with irrigation runnels (N24).  It is the ideograph or determinative in sp(3)t, “district, nome,” in hsp “garden,” and often appears in names of provinces.  Luwian hieroglyphs also include a grid, though the spaces are triangles and diamonds: TANA (which does not appear in my Latin dictionary, although a reader might provide a suggestion for the meaning).

Grids as found in Egypt (N24, upper left), Iran (M145~f, lower left), and Iraq (DARA4~c5, right).

Proto-cuneiform has more than one grid, neither of which is subdivided equally.  One represents a house or temple (E2~a), the other “blood; red, dark” (DARA4~c5).  Proto-Elamite contains a close analog of the first of these, although reversed (M145~bb), as well as a possible variant divided like an ice cube tray (M145~f).

Detail of patterns on Adire cloth (Yoruba area of Nigeria), showing grid or checkerboard at the bottom.

African Adinkra symbols also include a grid: ani bere a enso gya, “no matter how red eyed one becomes, one’s eyes will not spark flames” (Willis 1998: 78).  This expression and the symbol convey the virtues of patience and self-control.  Grids or checkboards are quite common around the world, as noted in a previous post.  Such a design is one of the apparently univeral entoptic patterns associated with altered states of consciousness (Lewis-Williams 2002: 128).  I see a checkerboard during some severe migraines, for example, a brightly colored pattern that "appears" with eyes open or closed.

Detail from a modern quilt of the simplest pattern, called the Nine Patch.
The top is pieced in a checkerboard arrangement of 3 x 3 x 3 squares (work of Rosa Sims).

Indus sign VIII 7 is another GRID (this one a QUILT in my earlier terminology, because it is subdivided 3 x 3).  This particular configuration appear only in Wells’ list, as W500, where ten occurrences are noted, eight from Mohenjo daro, one from Harappa, and one from Chanhujo daro.  Besides the previously cited parallels, there is the Adinkra checkerboard design kronti ne akwamu, “elders of the state” (Willis 1998: 122).  This symbolizes democracy, the duality of life, interdependence, and complementarity.

Detail of seal M-61: GRID (3 X 3 QUILT) / CEE / BI-QUOTES //
SINGLE POST / STRIPED TRIANGLE / TRIPLE TRIANGLES / POST.

Our final sign is CUP IN TRIPLE BRICK (VIII 8), also known as KP266 and W524.  It is a singleton from Mohenjo daro (M-952).  In form it is a square or rectangle with one dividing vertical stroke and a horizontal stroke that divides only one side.  There are few parallels.  Luwian hieroglyphs include a square with a central circle, meaning unknown.  Proto-cuneiform includes a rectangle with one vertical and a “C” shape that reminds me of the flap of a large envelope (GA2~a4).  It came to mean “box, basket; house,” and other things.  A similar sign without the “c” but with a tall, thin rectangle enclosed is |GA2~1 x GISZ@t|.  In other words, it is the same same as the previous, but now enclosing the symbol for “tree, wood” (and thus a wooden house?).

Detail from seal M-952 with partial inscription: CUP IN TRIPLE BRICK / VEE IN DIAMOND / (broken).

In some inscriptions of Old Chinese, the presence of the ancestor in the shrine is depicted with a triangle or semi-circle, and on occasion this is subdivided into three sections (Wieger 1965: 372).  It is supposed to indicate the ancestor’s sight, according to this author, and may be combined with other characters or symbols that also represent the ancestor (such as footprints and/or the character tian1).

Old chinese inscription showing the ancestor (anthropomorphic figure) in the shrine (enclosure).  The
subdivided triangle indicates the ancestor's sight, as does the large eye just beneath it (Wieger 1965: 372).

Luwian includes a square glyph enclosing a small, centered circle (enumerated 255).  Unfortunately, its meaning is not known.  In the rock art of North America, similarly, there are many motifs with a rectangular form that is subdivided asymmetrically.  These too are uninterpretable at present.

In modern times, quilts are typically made from square or rectangular elements, combined in various ways.  Some of the patterns have a form roughly similar to that of the Indus VIII 8, combining large and small rectangles with enclosures.  As an example, here is an illustration of a floral pattern.  I want to say it is called Grandmother's Flower Garden, but I think perhaps I have confused the name of a different quilt pattern with this one.  If any of my readers know the correct name, let me know.

Grandmother's Flower Garden (?) quilt, composed of squares that contain angular flowers (themselves made up of both squares and triangles), smaller squares, and smaller rectangles (work by Rosa Sims).

A Triangle and a Couple of Cups

The TRIANGLE is III22 among the Indus signs, also KP207.  It does not appear in either Wells' list or that of Fairservis, apparently because it appears only on pots.  I found one instance where it was painted (Rhd-226) and six incised occurrences (Rhd-80 through Rhd-86).  All of these occur at Rahman-Deri so far as I can tell.  One other instance also appears on a pot, or rather inside, although it is part of a more complex design.  It is incised with double lines, with what may be a rectangular element above.  This upper element is difficult for me to make out due to the flaring rim of the pot.  The triangle is clear in the photograph, however (H-996).  The other triangular signs, found at other sites, are bisected by a vertical line, striped horizontally, or both, and these do appear in the lists of other scholars.  Since they contain more than three strokes, I will discuss them later.

Decorative triangles on a pot produced by Brazilian natives

There is one simple triangle in Egyptian hieroglyphs, a tall and thin example (M44).  It represents a thorn.  A more common triangular glyph contains another smaller triangle inside, at its base (X8).  This is a conical bread loaf used in the verb "to give."  In hieratic writing the inner marking disappears so that it merges with the previous symbol.  Another triangular glyph is a little broader and has a thin rectangular base (O24).  This represents a pyramid.  Finally, a roughly triangular reed boat or float, used in fishing and in hippo hunting, with several additional details, provides one more distant parallel for the Indus sign (T25).

In Old Chinese, the simple triangle is ji2, "notion of union, of assemblage, of a junction of different elements, represented by three lines" (Wieger 1965: 45).  This character is not seen alone now but appears in compounds such as da2, "vetch, pea, vegetables whose boughs are joined, get entangled" (op cit.).  This explanation is one of those where I suspect Wieger's imagination may have gotten the better of his scholarship, but the latter character does seem to have the element ji2 in it, surmounted by two trident-like elements representing vegetation.  At the bottom is the mouth-like element which would ordinarily lead one to expect one portion to be a phonetic clue.  This does not seem to be the case, though, since ji2 and da2 are pretty different.  The latter is reconstructed as something like *tap in Early Chinese (Pulleyblank 1991: 69).  It's a bit difficult to discover the other since it has disappeared from the lexicon, but most words with that pronunciation are reconstructed as *ki or *tsi at a slightly earlier stage and from there to such Early Chinese forms as *gij, key, tsej, giajk, kejk, giajk, dziajk, tsik (1991: 138-139).  In other words, these two words weren't any more like each other in the past than they are at present.

Luwian hieroglyphs make use of a triangle for the syllable su.  Very early proto-cuneiform had a triangle also, the apex upward as all these previous ones did, which meant "to make, build" (Schmandt-Besserat 1996: 77).  Another triangular sign was positioned with the apex downward and had an additional horizontal line near the "base" (which was not the base but the top).  This represented a particular type of bowl with a beveled rim, many of which are found in archeological deposits.  This sign came to mean "bread," which is NINDA in Sumerian (1996: 73).  However, early on, the sign probably did not have precisely this significance, instead having some more administrative connotation having to do with the distribution of rations to workers.  At a later stage of proto-cuneiform, where signs are positioned horizontally by convention, a triangle with the apex to the right is KAK~a.  This sign came to mean "peg; nail; bone, joint, knee."  Yet another triangle is known as ZATU659 as its apex is placed to the left.  Presumably it meant something different from any of the others, but just what is unknown.  Proto-Elamite also has two different triangles, a long and thin one with the apex to the left (M106) and a broader one with the apex to the right (M129).

The triangle appears in rock art in North America also, not surprisingly.  Two appear in the Texas collection below a zigzag and between anthropoid figures (Newcomb 1996: 82, Pl. 43, no. 5).  The one on the right has the apex to the right; that on the left seems not to be equilateral and has the longest side on the left.  Another appears amid areas of cross-hatching (1996: 91, Pl. 51, no. 2).  Its outline is open on the right side, but whether this is intentional or due to weathering, I cannot say.  A third appears with the apex downward alongside a number of tally-like strokes (1996: 125, Pl. 82, no. 6).  The collection from Nevada and eastern California cites 19 occurrences of the triangle, including the following three (Heizer and Baumhoff 1984: 111, fig. 48b; 121, fig. 58c; 133, fig. 70b).  In the first of these, the apex is upward while in the last two, the apex is downward.

The next sign appears only once and is not listed by most researchers.  Its clumsy appellation is QUOTE BY QUOTE, its number III23.  It occurs on one pot shard, K-93 from Kalibangan, between FISH UNDER CHEVRON and FOOTED STOOL.  It looks like a short vertical line or SINGLE QUOTE above a low SLASH above another SINGLE QUOTE.  Thus, it might better be analyzed as a grouping of two independent signs, the PINCH plus a SINGLE QUOTE placed unusually low.  There are instances where other smaller signs are grouped in such a manner.  The BI-QUOTES occasionally are placed above the beginning of a row of apparent numerals, for example.  The less said about this peculiar "sign," the better.

The CUPPED POST is the next sign, III24, not separately listed by Koskenniemi and Parpola, W301, and Fs I-6.  I assume that it has no KP number because it was taken as a variant of the CUPPED SPOON or CUPPED Q-TIP as I originally called that sign (KP316).  Wells groups together the variants of this CUPPED POST (a "U" shape with a line inside that is not attached) and what I term the DUBYA where the "U" shape has a bisecting line inside that is attached.  He does not seem to treat the corresponding "V" shape with a bisecting line.  He states that his 301 has a frequency of 77, as follows: Mohenjo daro 54, Harappa 13, Lothal 4, Kalibangan 1, Chanhujo daro 3, Chandigarh 1, Rakhigarhi 1.  I seem to see a couple more, an additional one by his definition as Chandigarh, a "c" variant, and one his "c" variants as well at Amri.

Note triangle at top, indicating presence of Old Chinese ancestor;
note U shaped "mouth" element at bottom left (Wieger 1965: 371)

Fairservis actually agrees with the KP team, listing only the CUPPED SPOON formally.  Under the "comments" section for this sign, though, he shows the CUPPED POST.  So he, too, considers it a variant.  Apparently both are, to him, representations of mortar and pestle, signifying "hundred, a term used to mean many or innumerable" (1992: 170).

Egyptian hieroglyphs present only distant parallels in this case.  There is a bowl for incense with a sort of bent triangle at the top representing smoke coming out.  It is not similar in form but might possibly provide a hint of what the Harappan symbol could be depicting.  In the same vein, there is glyph W11.  This has three diffferent functions: (1) it is a ring-stand for jars, occasionally used as an ideograph or determinative for the word "ring-stand"; (2) it is a red earthenware pot used as a phonetic glyph for the g sound, its most common use; (3) very rarely, it is substituted for O45 during the 18th Dynasty, which should be domed building, in the expression for the king's harem.  From all this, if we wish to use Egyptian as an analogy, the vertical line in the Indus sign may be viewed as in the CUP, rising from the CUP, part of the CUP, or a decorative or symbolic element upon the surface of the CUP.  Since there is considerable variety in the way the POST and the CUP are shown on tablets, seals, and pots, we should have plenty of ammunition for whichever of these views we choose to hold (and plenty of dirty data to deal with as well).

Old Chinese has a "U"-shaped symbol that also has a line incorporated into the same character.  However, the line is horizontal, crossing the posts of the "U" rather than passing down the center as in the Indus sign.  This is nien4, which means the same thing as er4-shi2, "twenty" (Wieger 1965: 70).  "The tens added one to another.  In composition, it often means a multitude."  Nowadays it is written as two verticals with a crossing horizontal.  In other words, the "U" is no longer there.

Luwian has three different possibilities for an analogous symbol.  The syllable su (with an accent mark to indicate that it is a second way of writing su) is a type of "U" with a central post.  This is more complex, though, as the posts of the "U" itself have a bit of a wiggle to them at the tops.  The central post is also capped with a short, curved mark.  There is second "U" shaped sign, OCCIDENS, meaning "west," in which the central post is curved at the top like a shepherd's crook.  A third sign is squared off at the bottom, DOMINUS, meaning "lord."  It, too, is bisected by a vertical line.  It resembles an elongated "E" that is lying on its back with its prongs in the air.

Proto-Elamite has a horizontally oriented sign that somewhat resembles the CUPPED POST.  As is typical for this proto-writing system, the sign is more angular than the Indus sign.  It begins with something like a "greater than" mark: >.  On the left, horizontal lines are added to each end of this to form the outsides of the "cup," with a third line in the center for the "post."  In this case, the "post" does attach to the "cup."  There are two versions in proto-Elamite (M290~c and M-290~d; see http://cdli.uncla.edu/wiki/doku.php/linear-elamite ).  Thus, strictly speaking, it is more similar to the DUBYA sign in the Indus script.

There are instances of a CUPPED POST in the rock art of North America.  In Texas, this is rounded only and the central stroke is attached (Newcomb 1996: 131, Pl. 87).  Further west, both rounded and vee-shaped types occur and I have seen versions in which the central stroke is attached and others where it is not attached (Heizer and Baumhoff 1984: 143, fig. 80b and d).  In each instance cited, the form is upside-down compared to the Indus sign.  These, like the "tridents," may be bird tracks.  They are found in Australian rock art as well, oriented horizontally, at the Rockholes and Panaramitee Hill, Panaramitee Station (Flood 1997: 111, L section; also 101, Panaramitee North, Olary region, South Australia).

Wells divides the CUPPED POST into three variants: (a) short "cup" with tall unattached "post"; (b) tall "cup" with tall unattached "post"; and (c) tall "cup" with tall attached "post."  I see at least three variants also, but do not include Wells' "c" variant here, which I term the DUBYA.  My variants include: (a) "cup" half the height of the "post" (M-997, M-985, M-945, M-932, M-785, M-757, M-276, M-248, M-118, M-117, M-100, M-86, M-54, and M-50; H-175; C-9); (b) "cup" 3/4 the height of the "post" (M-928, M-860, M-814, M-797, M-754, M-714, M-426, M-286, M-122, M-44, M-4; H-206; L-10, L-11, L-144(?), L-146, L-191(?)); (c) "cup" and "post" of equal height (M-1364, M-1355, M-365, M-359, M-355, M-1226, M-1191, M-1067, M-1057, M-957, M-863, M-677, M-670; H-24, H-25, H-28, H-136, H-210B(?); L-85, L-115, L-183; C-7, C-35; Rgr-2).  There are a few misshapen ones: M-72, M-128, M-135, M-303.  Some of those that Wells lists are what I call DUBYA:  M-864, Ch-1, along with a couple more, including Ch-2 and Amri-1.  There are two that I can't be certain of.  They may be the DUBYA but may be the CUPPED POST: M-1299, M-1328.

Replica of Indus tablet with CUPPED POST (2nd from right)

Another sign begins with a "cup" shape but is much rare.  It is the CUPPED ROOF, III25.  This appears only as KP318, not in either Wells' list or in that of Fairservis.  The only certain appearance is in L-36 from Lothal, although an impression from the same locale may show it (L-183).  A similar symbol seems to occur at Mohenjo daro on a particularly obscure tablet (M-481 B) where there seem to be striped at the bottom.

The Indus sign is somewhat reminiscent of the Egyptian well full of water (N42), a "U" shape with a horizontal line near the top.  The Old Chinese kou3, "mouth" is nearly identical in form, though not in meaning (Wieger 1965: 180).  Luwian hieroglyphs include a half circle with the flat side up, CAELUM, meaning "sky."  Underneath the flat top there is another horizontal line.  Between the two horizontal lines are four very short vertical strokes.  Thus, this glyph is considerably more complex than the Indus sign.

There is a similar motif in the rock of western America except that it is upside-down compared to the Indus symbol (Heizer and Baumhoff 1984: 155, fig. 92j; 161, fig. 98a).  It appears twice.  The first occurence cited has an extension of one prong of the "U" while the second has a horizontal line closing it at the bottom rather than a curved line inside.

Our last sign for today is the dotted circle or, to be consistent with other terms, the CIRCLED DOT, III26.  It appears in a very round form with a round dot in the center, in a pointed oval form with a short vertical in the center (i.e., a quote in the center), and as a very thin, pointed oval with a central quote or dot.  It was formerly KP366, W346, and Fs N-9.  Fairservis considers it the sun, meaning "to set, to rise, do, obtain."  When doubled, it means "boats," he suggests.  All these meanings are homophonous or nearly so in Dravidian languages, he tells us.  He seems to see the dot or quote inside as a mark of the genitive/possessive case (1992: 180).  It does seem odd to me, though, to have a single symbol for both rising and setting.  Those appear to be rather different from my point of view.  But what do I know?

Apotropaic jewelry to repel the Evil Eye: a modern CIRCLED DOT

The Egyptian hieroglyph for the sun is a circle with a smaller circle inside, ideally (N5), used as an ideograph or determinative in the word r', "sun, day."  I would like to point out that this is not an extension of the use of one glyph for two different words in Egyptian.  In that language the one word meant both things, even though that is not the case in English.  Some writers on this topic fail to get this point.  Another Egyptian circular glyph has a horizontal line through it (N9).  This is the moon, an ideograph or determinative in the word psdntyw, "new moon festival."  A third circular glyph has two little quotes in it (O48), a version of a prehistoric building found at Hieraconpolis (dating to the 18th Dynasty, other versions of the glyph being used at other times).  This sign is used as an ideograph or determinative for that city.  A fourth circular glyph has many dots inside (O50).  It represents a threshing floor.  Finally, X6 has a squarish or ROOF-shaped mark at the bottom and is one of many representations of a bread loaf.  (By the way, when I say that a glyph is an ideograph or a determinative, I am not confused and neither were the Egyptians; the sign is sometimes one and sometimes the other, depending on a variety of factors that are tedious to go into.)

Old Chinese once had a circle with a dot or line, similar to the Egyptian sun and moon.  Both were ri4, "sun", the 72nd radical (Wieger 1965: 311).  As an aside, this word also means "day" in Chinese.  The old style writing of liang2, "gift" includes this element, at one time written as a circle with a horizontal line across the middle with one dot above the line and a second dot below the line.  According to Wieger, this was intended to show the meaning of a gift in the sense of a talent as something coming down from heaven: "gift....original qualities, nature, inborn, good" (1965: 194).

Luwian hieroglyphs include a circle with a smaller circle inside, representing the syllable sa4.  Proto-cuneiform also has a circle in a circle, conventionally transcribed as |LAGAB~a x LAGAB~a|, which may represent a lamb.  Another circular symbol has what I would describe as a "stacked eight" (two neat columns of four short strokes each) inside.  This is GUG2, which came to represent a type of cake made of pressed dates.  Proto-Elamite has a circular symbol with another smaller circular element inside, both of which are impressed, not incised.  This may be a numerical sign.

Motifs of dotted circles and concentric circles are both common in the rock art of North America.  The circled dot appears in Texas (Newcomb 1996: 101, Pl. 57, Pl. 58).  Here it is well rounded.  In the collection from the Far West, the authors report 139 concentric rings that include 2-8 rings and a total of 60 occurrences of the circle and dot motif (e.g. Heizer and Baumhoff 1984: 104, fig. 41b & j; 141, fig. 78c).  Some of these are rounded, some more oval.  Dotted circles also appear in great number engraved on rocks in Australia (Flood 1997: 185, Wharton Hill, Olary region, South Australia; 197, Frank Creek, Mount Isa, Queensland; 238, Trial Harbour, Northwest Tasmania).  In fact, the "cup and ring" motif as the dotted circle is often called, is one of the oldest motifs found in rock art.  It appears to predate the appearance of Homo sapiens in some areas.

The dotted circle and concentric rings also occur as entoptic forms.  I "see" both during migraines and when my eyes are closed or in a darkened room when I press on my eyelids.  It seems to me that a small circle appears, it grows and develops a dot in the center.  This dotted circle grows and becomes a circle with a smaller circle in the center.  This continues to grow, developing a new dot in the center.  At this point, the outer circle grows to the point that its outer margin begins to fall beyond the limits of my inner "vision."  But the growing of these circles within circles can continue for some time.  I mentioned in a previous post the theory that the origin of art is to be sought in shamans depicting their visions in stages of trance (Lewis-Williams 2002).  If the theory is correct, the circled dot and the circled circle originally appeared in prehistoric rock art because a shaman saw them in the early stages of a trance, then.

Lewis-Williams suggests that the shaman goes through three stages of altered consciousness.  In the first stage, he (or she) sees entoptic forms and perhaps hears buzzing sounds, feels prickling or the skin, and so on.  These sensations have been replicated in the laboratory by subjects taking hallucinogenic drugs.  I can attest to many of the same things during severe migraines, including prickly sensations, hearing buzzing, and seeing entoptic forms (although I quibble with the list of 15 specific shapes). 

The second stage is more problematic to my thinking.  In this stage, subjects are supposed to try to make sense of their peculiar sensations.  Unaware that I was supposed to do this, I failed to have such a stage during any migraine.  But Lewis-Williams describes this phase for us: ""subjects try to make sense of entoptic phenomena by elaborating them into iconic forms, that is, into objects that are familiar to them from their daily life....The brain attempts to decode these [entoptic] forms as it does impressions supplied by the nervous system in an alert, outwardly-directed state.  This process is linked to the disposition of the subject.  For example, an ambiguous round shape may be 'illusioned' into an organge if the subject is hungry, a breast if he is in a state of heightened sexual drive, a cup of water if the subject is thirsty, or an anarchist's bomb if the subject is fearful" (2002: 127).  Well, perhaps if I'd had a lab assistant or an anthropologist standing by, taking notes on my illusions, I might have had made such an attempt at interpretation of my little "fireflies."

Stage 3 occurs after the subject has gone through what is often described as a vortex or tunnel, something also associated with near-death experiences.  At this point, the subject hallucinates iconic images that do not derive from the optic system itself but from the mind.  Alongside these dream-like images, entoptic shapes may also appear.  Lewis-Williams suggests that odd things may now happen to these basic shapes and to the subject:

"By a process of fragmentation and integration, compound images are formed: for example, a man with zigzag legs.  Finally, in this stage, subjects enter into and participate in their own imagery: they are part of a strange realm.  They blend with both their geometric and their iconic imagery.  It is in this final stage that people sometimes feel themselves to be turning into animals and undergoing other frightening or exalting transformations" (2002: 130).

Peculiar as it may seem, between the severe pain of migraines (which sometimes lasted for days and even weeks at a time) and powerful medications, I occasionally passed through a vortex similar to what is described and felt transformed.  However, I did not detect a second stage since there was no one to interpret the experience to.  I was alone with the pain and images and did not consider the possibility that there was a meaning until much later, looking back on the experience.  Thus, I tend to think that researchers only find THREE stages because of something I mentioned in a previous post.  Our culture places considerable emphasis on three as a mystical, magical number and has for untold millennia.  We expect things to come in threes.  So we tend to find things in threes, even when they aren't really there.  I think there are only TWO stages of trance in reality.  There is only a third, middle stage if we decide to put it there.