Tuesday, December 28, 2010

An Attached Trident, a Five-Pronged Fork, and Odd Signs between

The first Indus sign considered in this discussion is a set of three tall vertical strokes, with a trident angling off on one side.  I call this THREE POSTS WITH ATTACHED TRI-FORK (VI33).  Elsewhere, this is cited as KP148 and W222, but not in Fairservis.  Wells notes a single instance (M-898).  A sign previously discussed, shown only in Wells’ list, may be a second occurrence, though.  Sign W182 is what I term DOUBLE QUOTES ON SLASH-BELTED AITCH (V14).  I appears at Mohenjo daro also (M-840).  In the latter inscription, four “posts” occur to the left of an enigmatic element and three “posts” to the right of the same element.  Wells apparently takes this sequence as THREE POSTS / DOUBLE QUOTES ON SLASH-BELTED AITCH / TWO POSTS.  But we could interpret it just as reasonably as FOUR POSTS / THREE POSTS WITH ATTACHED TRI-FORK. 
(smoothing of image and color added).

There are other Indus signs with an attached trident-like element (see TRI-FORK III13 for a full listing).  But no other apparent numerals occur with it.  Similarly, a trident-like element frequently appears in Old Chinese, usually representing either a schematic hand or a plant (Wieger 1965: 119-120 on the hand; 696 on the “sprout”).  But it does not occur in compound characters with numerals, so far as I can determine.


The closest parallel to this Indus sign (VI33) comes from Luwian hieroglyphs.  Three posts with a backslash attached to the one on the far right form the glyph for tara/tari.  It is possible that the sounds represented derive from an underlying Luwian word “three,” since most Indo-European cognates are phonetically similar (derived from Proto-Indo-European reconstructed form *trei-, Watkins 2000: 93).  Thus, the three verticals may represent the numeral and the additional stroke may indicate the pronunciation of that number word without its accompanying meaning.
Proto-Elamite also includes a symbol comprising three parallel strokes, horizontal in this case, plus an additional element.  The addition here is a half-circle attached beneath the central line, making it take on the appearance of the letter “P” lying on its face.  The meaning of this sign is unknown (M009~e).
Today’s second sign is FAT CHEVRON (VI34).  It is an upside-down “V” shape, i.e., a chevron.  Unlike the simple, two-stroke version, this one is made up of two thin, overlapping rectangles.  It also appears as KP197, W415, and Fs K-1.  Fairservis thinks that it represents a carpenter’s square or rule meaning “(linear) measure; to unite, join.”  Wells notes a total of 20 occurrences, eight from Mohenjo daro, 11 from Harappa, and one from Lothal.  My count differs: 14 from Mohenjo daro, perhaps 19 from Harappa, two from Lothal, and perhaps two from Banawali).

Inscription on tablet H-808B (reading from right): SINGLE POST / VEE IN DIAMOND / BI-QUOTES /
WHISKERED FISH / FAT CHEVRON / POT (coloring added and image smoothed a bit).

There is a very similar glyph in Egyptian typically written at a different angle (O38).  This glyph resembles an upside-down “L” due to its position.  It apparently depicts the corner of a wall as it functions as a determinative in “corner, angle.”  The Indus sign might also be compared with Egyptian glyph T14, a throw-stick or club, sometimes used in names of foreign peoples.
In Old Chinese, there is no thick-armed chevron, but the simple chevron occurs twice, more or less.  The most basic of these, supposedly, represents a suspended object (Weiger 1965: 42).  This does not seem to occur individually, but only in complex characters.  The other version has an additional bit of vertical at the peak, ru4, “to enter, put in....The character represents the penetration of roots into the earth” (1965: 50).  It is now the 11th radical.
In proto-cuneiform, a symbol much like the Egyptian glyph of the corner appears in the position of our “L.”  This is UR~b, which came to mean “dog; young man; servant; warrior.”  The same sign does not occur in proto-Elamite, but an angled type does, which is striped (M316~e, a “fat less than” form).
We now come to a rather peculiar sign that only Koskenniemi and Parpola observe, the HALF FAT KAY (VI35).  This is similar to the letter “L,” with a thin rectangle rising from the corner of the angle like a thick slash.  Another short line, angled like a backslash, descends from this thick element.  This much resembles our letter “K,” with only one bit thickened.  There is one more line though, a horizontal stroke at the bottom. 
Inscription on poorly executed (?) seal M-291: FAT EX (?) / DEE (?) / POT (color and smoothing added).

I have not observed this sign in the Corpus and it does not appear in the lists by Wells or Fairservis.  It might possibly be an interpretation of one of the many fragmentary signs found on broken seals, signs which I may have seen differently.  For example, seal M-1340 appears – to my eyes – to have part of the PRAWN at the break on the left, followed by ZEE / CROSSROADS EX / POT / SINGLE QUOTE / PRICKLY CORN HOLDER.  The right-hand portion of the PRAWN, all that is preserved, has some features in common with the HALF FAT KAY.  It lacks the horizontal stroke at the bottom, though, and has a couple of extra lines that do not appear in the sign as depicted in the KP list.  I think it more likely that sign VI35 is a different reading of the first symbol on the left on seal M-291.  This sign to my eyes is a poorly executed FAT EX.  To this peculiarly proportioned sign we may compare the equally odd quadruped beneath the inscription.  This animal representation, assuming that's what it is, has been so simplified that I cannot tell what it is.  Cat?  Dog?  Tiger?  Cow?
As far as parallels go, I find only one and it is not particularly close.  In proto-cuneiform, one thick “less than” sign variant has a “ladder” extending horizontally from the angle.  This is |UR~a x KAR2|, a ligature of two signs, possibly a combination of their later meanings: “dog; warrior” + “encircle; illuminate.”  Would that make it an illuminated warrior?  Or is it an encircled dog?  More than likely, it is neither.
We now turn to three signs that are similar in the Indus corpus and that may be related.  First is FLAIL OVER FOUR QUOTES (VI36).  This takes the form of the number “7” tilted far to the left, with a row of short strokes below it on the left.  This version is not shown in the KP list or in Fairservis; it is W445.  Wells notes it as a singleton (H-143). 
Inscription from seal H-143: TWO POSTS / FAT EX IN DIAMOND / POT / SINGLE QUOTE /
The “flail” portion has been discussed before among the two-stroke signs (FLAIL, II13).  It resembles the Egyptian hieroglyph representing the adze from the Old Kingdom (U20), its angle more similar to the flagellum (S45).  The FOUR QUOTES sign has also been covered (IV1).  The closest parallel to this as a non-numerical sign appears in Luwian, for the syllable mi, perhaps related to the word mawa, “four.”  Luwian hieroglyphs do not include a ligature between a flail-like element and these four strokes, but there is a syllabic sign that is reminiscent of the Indus sign.  The glyph comprises an upside-down “7” with three short slashes attached on the right, for ta5.  Here, similar elements are arranged in a completely different manner.

CIRCLED TRI-FORK / POT-HATTED BEARER (image smoothed and colored).

The next Indus sign is quite similar, AY OVER THREE QUOTES (VI37).  The “7” element now has an additional stroke, not quite horizontal, joining the two angled lines.  This forms an element resembling the letter “A” with one leg longer than the other.  Below the short leg are three short verticals or “quotes.”  Again, only Wells enumerates this (W437).There are five occurrences of this sign, two from Mohenjo daro and three from Harappa.
The most similar sign elsewhere is now from Egyptian, in the ‘low (U13).  The glyph, unlike VI37, lies on its shorter leg, and the additional strokes – of which there are only two – rise at an angle from the longer leg.  Thus, as with the Luwian glyph cited for the previous sign, these are similar elements but arranged rather differently.  Even so, the Indus AY may represent a simple plow, as in Egyptian.
On the other hand, the sign that came to mean “plow” in proto-cuneiform is quite different in form.  One variant of this is laid out horizontally, while two others are vertically oriented (APIN~b horizontal, with APIN~a and ~c vertical).  This sign resembles the Indus PINWHEEL (ZEE) with an “A” attached at one end.  In addition, while the Indus AY might represent a plow, that is by no means certain.  Fairservis considers a more complex sign I term SQUIRREL to be a plow (my VIII46).
Regardless of what the Indus sign represents, it is interesting to note that the inscription containing VI36 almost duplicates one inscription containing VI37.  Both are from Harappa and are shown in the accompanying illustrations (H-143: TWO POSTS / FAT EX IN DIAMOND / POT / SINGLE QUOTE / FLAIL OVER FOUR QUOTES / CIRCLED TRI-FORK / POT-HATTED BEARER; H-146: SINGLE POST / AY OVER THREE QUOTES / CIRCLED TRI-FORK / POT-HATTED BEARER).  The last three signs in each case are very similar, the last two virtually identical.  This could be interpreted as evidence that FLAIL OVER FOUR QUOTES means the same thing as AY OVER THREE QUOTES.  It is not proof, however, as a single example is insufficient for that.

Another apparent variation on the same sign is our next consideration, AY ON TRI-FORK (VI38).  Again, the main element resembles the letter “A” with uneven legs, but this time the three short strokes beneath the short leg are attached to one another and to the “ay.”  Again, only Wells enumerates this separately and again it is a singleton (M-40).  Besides comparing the Egyptian and proto-cuneiform plow signs, we might also glance at variants of a proto-Elamite sign that resembles APIN (M056, M056~e, ~f, and ~g).  In these, the “A” element is upside-down and tilted.
Inscription detail from seal M-40: FIVE QUOTES / AY ON TRI-FORK / PINWHEEL (ZEE) /
 BI-QUOTES / TRI-FORK UNDER CHEVRON (smoothing and added coloring by author).
Wells observes a total of nine different signs that resemble this one: AY OVER FOUR QUOTES (W436), AY OVER THREE QUOTES (W437), AY ON E TRI-FORK (W438), AY ON COMB (W439), AY WITH ATTACHED QUAD-FORK (W443), AY OVER FIVE QUOTES (W444), FLAIL OVER FOUR QUOTES (W445), AND AY ON BACK E TRI-FORK (W446).  Those not discussed here will be included in later discussions of seven-stroke and eight-stroke signs.
The next sign is a very odd ligature, EX ON POST WITH ATTACHED LEG (OR ATTACHED EAR ON BACKSLASH).  I enumerate it VI39, the thirty-ninth of the six-stroke signs.  It is yet another of the many signs only found in Wells’ list, where it is W67.  Wells finds it to be a singleton (M-800).  There are other Indus signs with the small addition I term the “ear,” but none of these others places this element on a long post or slash.  Other signs appear to be varied by the addition of a single stroke, either vertical or diagonal, which I usually term a “tick.”  None of these others adds the “ear” to the “tick,” as in VI39.  In addition, the basic element of this sign, the EX ON POST, is itself a singleton (see III14).

Inscription detail from seal M-800: CORN HOLDER / EX ON POST WITH ATTACHED LEG / POTTED ONE /
CIRCLED TRI-FORK / CAGED BELTED FISH (smoothing and coloring by author).

The next two signs are among those I term PANTS for want of a good name, distinguishing these six-stroke versions as SEAMLESS PANTS (VI40) and RUMPLED PANTS (VI41).  Neither version appears outside of Wells’ list, where the first is W166 and the second W169.  Wells also notes a five-stroke version (W168), two comprising seven strokes (W161 and W167), two of eight strokes (W160 and W163), as well as nine-stroke versions (W162, W170, W165, W172).  Most are singletons, but four versions occur from three to seven times. 
Even if we lump them all together as variants of a single sign, the 11 variations occur a total of only 24 times.  Of these, 19 are from Mohenjo daro (divided among 9 variants), two from Harappa (two variants), one from Banawali, one from Chanhujo daro, one from Kalibangan, and one from Khirsara.  The SEAMLESS PANTS (VI40) appear at Mohenjo daro on a bangle (M-1631); the RUMPLED PANTS (VI41) on a seal from the same place (M-1117). 
Inscription detail from bangle M-1631: SINGLE QUOTE / CIRCLE / SEAMLESS PANTS /
POTTED ONE (coloring and smoothing by author)

Proto-Elamite provides a weak parallel, positioned horizontally (M059~d).  This sign includes three bent lines that parallel each other and are of the same length, joined at one end and at their bend.  The SEAMLESS PANTS among the Indus variants is the closest to this with two bent lines of the same length joined at the top.  Between them is a single shorter vertical at the bottom.  Other variations on the PANTS are less similar to the proto-Elamite example.

Seal M-1117 with inscription: VEST (?) / RUMPLED PANTS / DIAMOND BETWEEN DOUBLE POSTS (image smoothed and colored).  Note the unusual pose of the iconic animal, looking backward over its shoulder.

Old Chinese contains a character composed of five wavy lines.  It is yong3, “the unceasing flow of water veins in the earth....Abstracted meaning, duration, perpetuity, but not eternity” (Wieger 1965: 289).  The same lines inverted form pai4, “the idea is analogous; ramification of a stream” (1965: 289).  If the Indus sign is also intended to represent flowing water, it might actually indicate more than one thing with the different variations – rivers, streams, lakes, ponds, the ocean, irrigation channels?
The final sign to be considered here is another of the “forks,” the QUINT-FORK (VI42).  It takes the form of a “Y” with one prong longer than the other.  From the longer prong rise three diagonal strokes.  Thus, counting the two prongs at the top of the “Y” plus the three additional strokes, there are five “tines” to this “fork.”  Elsewhere, this is KP86 (“A” variant, with additional strokes on the right), W278, and the reverse is W265 (“B” variant, with additional strokes on the left). 
Inscription from broken seal H-657: (?) / FISH UNDER CHEVRON / DOT IN FISH / LAMBDA / QUINT-FORK (left branching).  I have completed the chevron but cannot guess the identity of the first sign.
Both Wells and Koskenniemi and Parpola’s list reverse the forms as they appear on the seals.  The actual right-branching variant is a singleton (M-749), while the left-branching variant occurs four times in my database (H-657, H-659, H-689, H-719).  Fairservis designates this sign (right-branching variant only) as E-2, seeing it as a stalk of grain.  He proposes the meaning “lunar month; order, line, row.”
The Egyptian glyph of the lotus flower resembles the Indus sign only in that the “prongs” or petals are all on one side (M9).  Plus there are five petals on this flower whose stem leans to the side.  A much better parallel appears in proto-Elamite (M039~a, M-039~b).  The “a” variant is a “comb” with “teeth” to the left, and a long horizontal stroke attached on the right, at the top.  In the “b” variant, the long horizontal is attached at the bottom and the “comb” element is angled.  In both variants, the “comb” has five “teeth.”  Unfortunately, meaning is unknown.
In the Mixtec proto-writing of Central America, human footprints are often represented in a schematic fashion which somewhat resembles the Indus QUINT-FORK.  The basic footprint is something like a “7” shape, though the angle is not as sharp and is usually curved.  Above this, there are five dots or five short strokes, representing the print of the toes.  Lines of footprints indicate a journey of an individual to another place and visits to other individuals.
While the human footprint appears elsewhere, as in the rock art of the Americas, of Africa, and of Australia, it does not take such a schematic form.  However, the tracks of various animals also appear in such art and one of another of these might inspire a symbol such as the QUINT-FORK.  The wolf, fox, and coyote have just four toes, typically, but the bear has five like the human.  It is also possible that a schematic footprint or track might be made of an animal such as a wolf, but purposely given an additional toe in order to represent a supernatural Wolf deity or Wolf ancestor.  Then again, it is just as likely that Fairservis is correct in seeing the QUINT-FORK as a type of plant.

Parker, Philip M. 2008. Webster’s Luwian-English Thesaurus Dictionary. San Diego: ICON Group International (see online www.websters-online-dictionary.org )
Watkins, Calvert. 2000. The American Heritage Dictionary of Indo-European Roots (2nd edition). Boston: Houghton Mifflin.

Saturday, December 25, 2010

Stools, Tables, an Asterisk, and a Rake

I begin this post with a sign found only in the list of Koskenniemi and Parpola (KP241).  I call it CAGED EX since this list shows it to have the form of an “X” shape surrounded by four short marks (upper right, upper left, lower right, and lower left).  In my own list it is VI26, indicating that it is the twenty-sixth of the six-stroke signs.  I am not actually certain that it exists, as I have no mention of it in my database of inscriptions, which are based on my readings of the first two volumes of the Corpus. 
The enigmatic seal M-787: VEE IN DIAMOND (?) / BI-QUOTES / CAGED EX (?) / BEARER
 (some smoothing of image and false color added by author).

It is possible that the sign is present on one seal from Mohenjo daro that is both broken and highly abraded (M-787).  The upper left-hand corner is broken off, removing the top of the first sign in the inscription.  The pointed, angular portion that remains at bottom of this sign suggests that it may have been VEE IN DIAMOND, a common variable appearing in prefixes.  It is followed by the BI-QUOTES, the most frequent of the constants in prefixes.  This much is reasonably clear, despite some abrasion.  The next sign is too abraded for me to be at all sure of its identity.  The two left-hand marks of “caging” appear to be present, and possibly that on the lower right.  The central portion could be an “X” shape, although that identification is doubtful.  The final sign is the BEARER with straight arms.
A slightly different sign is also possible here, and perhaps a little more likely, considering the poor state of seal M-787.  Among the five-stroke signs, I discussed the EX UNDER TABLE (V25).  This has the same central element, the “X” shape, as the proposed CAGED EX.  The center of this sign, of course, is the least visible portion on the seal, as seen in the first illustration.  If the top horizontal of the TABLE was originally there but is now too abraded to see, that might leave the “legs,” making them appear to be the two upper marks of caging.  Against this, it is the two left-hand marks that I see, and the rest of the sign more resembles one of the birds than an “X” to my eyes.  Be that as it may, we will assume that a CAGED EX is present somewhere in the Corpus.
A motif similar to Indus caging, detail from textile made by Native Americans of Peru (reference below).

Although the CAGED EX seems to be a simple symbol, I have not seen another example
of it, either in Indus inscriptions or among the symbols of other cultures.  Caging with four dots is fairly common in the Indus script but rare elsewhere, so far as I can tell.  An “X” shape is common around the world, but not in combination with bracketing.

Distorted image of a detail from seal C-6: SPEAR / JAY / SINGLE QUOTE / FISH BETWEEN PARENTHESES / POT / MAN // DOUBLE CARPET RAKES / CIRCLED VEE (smoothing and false color added by author).

The second Indus sign I will discuss resembles the letter “T” with some additions.  On either side of the horizontal stroke at the top, there is a short vertical mark.  A small triangle sits in the center of the horizontal as well.  I call this sign a CARPET RAKE as it resembles a rake I once owned that was intended for “combing” shag carpets.  It also somewhat resembles one of the BEARER signs of this script (CHEVRON HATTED BEARER), but with the arms raised and without legs.  It occurs elsewhere as KP94 and W245.  Fairservis makes no mention of it.  Wells gives a total of three occurrences, two from Mohenjo daro and one from Harappa (M-3, M-192; H-90).  I see an additional occurrence from Chanhujo-daro which is doubled (C-6).  The absence of the last from Wells’ enumeration is due to his predilection for classifying cases of doubling as independent signs.
The CARPET RAKE is not duplicated precisely anywhere else that I have seen, unlike the ordinary RAKE.  Distant similarities do exist here and there.  Among Egyptian hieroglyphs, there are two representations of a small table or cult stand (a platform on one leg) on which sit two loaves of bread and a jug.  The first version of this glyph shows three items on the table, a tall loaf on one side, a jug without handles in the center, and a small, round loaf on the other side (R1).  This is an ideograph or determinative in h3(w)t, “table of offerings.”  The second version is a schematic variant of the previous glyph with up to five strokes (vertical lines with bent tops) on the table (R2). 
In connection with an interpretation of the Indus sign as a person’s arms, hat, and trunk, one might also compare the ideograph for k3, “soul, spirit” (D28).  This Egyptian glyph represents two conventionalized arms upraised, recalling the prongs on the right and left sides of the top of the Indus sign.  But there is neither a head/hat nor a body of any kind in the glyph, while the Indus sign has both.  The final hieroglyph to compare resembles a triangular hat with arms coming down and holding an oar or paddle (D33).  The glyph is ideographic in hni, “to row.”  It perhaps resembles the CARPET RAKE only in my mind.
Proto-cuneiform also contains an analog to the Indus sign with |SIG x 1 (N57)|.  This resembles a horizontal version of the CARPET RAKE in which the “arms” and “hat” are curved rather than angular.  Here, the curved portion means “evening,” probably based on a representation of a crescent moon.  The single horizontal on the left is a numerical symbol, with the meaning “one.”  Proto-Elamite presents a more angular analog, also horizontal {M036 + 1 (N14)}.  The “arms” of this symbol are on the left, the opposite of the proto-cuneiform sign.  Instead of an angular “hat,” there is a round impression between the “arms.”  The meaning of this ancient Iranian symbol may or may not be related to “evening.”
Examples of
with "mountain."

Old Chinese has no precise parallel, but two less than perfect examples.  The first is based on a “U” shape, with a chevron inside from which extends a vertical post.  This is shan1, “mountain” (Wieger 1965: 208).  Today, this character is angular, a horizontal line with three verticals attached.  There is one on each end and a third in the center which is taller than the other two.  The character is the 46th radical.  Obviously the lower vertical of the Indus CARPET RAKE is missing from this Chinese parallel.  My second example is a curved trident in form, the old character che4, “a plant that sprouts from its grain; the minimum of a plant...often used as a symbol...to represent any object” (Wieger 1965: 199).  It is now the 45th radical, its top portion identical to the modern “mountain,” but with a central “post” that tends to curve toward the bottom.  In the accompanying illustration, a more complex character includes the "mountain" element at the bottom (yue4, "mountain peak; wife's parents" {Keightley 1978: 218}).
Sign VI28 is another BOWTIE, this one with two stripes, one on each side (cf. the unstriped IV10 and the single-striped V61).  It appears elsewhere only as W464, where it is classified as a singleton (M-855).  However, I see it on two more objects from Mohenjo daro (M-119, and below the markhor on M-1129), in one inscription from Lothal (L104), and perhaps on a bangle from Kalako-deray (Kd-8).  It also resembles the central element of a more complex design on a Post-Harappan seal from Pirak (P-25).  Even if all of these are classified as the same sign (which one may not wish to do), there are still very few altogether, certainly too few for statistical analysis.

Detail from seal M-119: STRIPED BOWTIE / CEE / TWO POSTS / BLANKET (?) /

A STRIPED BOWTIE symbol is not particularly common outside the Indus Valley either, in my observations.  Proto-cuneiform includes one rotated 90 degrees in ZAG~c.  The internal stripes on this symbol are different from those on the Indus sign, though, so this is not an exact match.  The Sumerian descendant of this sign means “boundary, border; sanctuary, shrine” and is a measure for fish (as well as a few other things).  One proto-Elamite sign is also shaped like the BOWTIE and has some stripes inside (M286~b).  There are four stripes in this symbol, all on one side, and the whole thing is rotated as in proto-cuneiform.
The next sign is similar to the previous but the stroke that would close one side is broken in the center.  I term this FOOTED STOOL, and this particular version also has a vertical line down the center.  This yields a full but clumsy descriptor, FOOTED STOOL WITH MID POST (VI29).  It is also KP231 and W458 but Fairservis does not note this variant.  Wells cites a total of two occurrences, both from Mohenjo daro (M-632 and M-148).  I see it on one tablet from Harappa also (H-953).  The “feet” are on the left side on the two seals from Mohenjo daro, on the right on the tablet from Harappa.  This is as expected, since seals were used to stamp things and tablets apparently were not.
Detail from seal M-148: FOOTED STOOL WITH MID POST / CIRCLED VEE / POT (color and smoothing added).

A single analogous form occurs in proto-cuneiform as GA’AR~b1.  This is actually another “bowtie” shape, but with the central vertical post found in the Indus sign.  In proto-cuneiform, the sign indicates a commodity made from dried, grated cheese.
A similar sign in the Indus script is VI30, a FOOTED STOOL WITH TICK.  Here, there is a short diagonal mark attached to one of the “legs” of the “stool” rather than a tall post crossing it.  This is also KP230(b) and W448, again missing from Fairservis.  Wells notes a total of 14 occurrences, nine from Mohenjo daro, two from Harappa, and three from Lothal.  I found 13 from Mohenjo daro, four from Harappa, three from Lothal (as Wells says), and an additional one of unknown provenance (a total of 21).  Either way, occurrences are insufficient for statistical analysis.
STACKED TWELVE / POT // FOOTED STOOL WITH TICK / PINWHEEL (color and smoothing added).

I have no parallels to cite for this sign other than the same one in proto-cuneiform noted earlier, GA’AR~b1.  The later Sumerian version indicates “powdered or finely grated sun-dried curd-cheese which can be stored and reconstituted with water or milk” (Halloran 1999: 69).  There is a compound in the same language, ga-ar3-gazi, meaning “cheese seasoned with gazi, a pungent spice – cassia or black mustard” (1999: 69)  If the Indus STOOL or FOOTED STOOL represents some type of commodity, the additional “tick” mark might indicate a variation on it, in a similar manner (but that’s a big “if”).
Another singleton among Indus signs is something of an anomaly, EN UNDER TABLE (VI31).  It also occurs as KP189 and W433 (missing from Fairservis).  The bottom portion, the EN, is a very short zigzag, resembling our capital letter “N” except that the vertical strokes lean to the right.  Over this is the TABLE, a square bracket with its “legs” hanging down.  The apparent ligature of these two elements appears only on bar seal M-367, as Wells notes.  Typically, symbols found beneath the TABLE element occur elsewhere as independent signs without the TABLE.  But the EN element does not appear alone, unless we view it as a variant of the EM WITH TICK, or other longer zigzags.  A less probable relationship might be with a sign inscribed on a broken pot rim, M-1587.  The photograph in the Corpus is not clear, and there could be one or even two instances of the EN on this object.  However, I think it more likely that these are occurrences of the BOWTIE.

(color and smoothing added; lower right corner missing on original).
When it comes to parallels, VI31 is again a singleton.  There are symbols in other scripts which are analogous to either the TABLE (see post on III6) or the EN (cf. the zigzag EM WITH TICK, V8), but not for the combination of these.  For example, Egyptian hieroglyphs include the common phonetic sign for the n sound (N35).  This is a zigzag, but one with a larger number of strokes than the EN.  There is also the sky element.  It occurs in two glyphs with something descending from it (N2 and N4).  But the sky does not appear over the water (i.e., n) as a single glyph.
In Luwian hieroglyphs, an element resembling the Indus GRAIN EAR appears beneath an apparent TABLE, both in the logograph SCRIBA, “scribe,” and in the phonetic glyph tu.  But no zigzag of any length occurs below this same “table” motif in Luwian.  In proto-cuneiform, there is a sign of unknown meaning that is vaguely similar to the Indus TABLE.  It is almost identical to the Egyptian glyph of the sky, rotated 90 degrees (ZATU751).  Hence, it is a poor analog to the Indus VI31.  Finally, proto-Elamite contains apparent variants on the Indus EN (M050~k, M050~m, and M050~n).  All have additional strokes resembling hairs and none occur within the long-legged proto-Elamite analog of the TABLE (cf. M106 + M288, where a long-legged “table” encloses a sign resembling the Indus VEST).
Detail from seal H-43: TWO POSTS / ASTERISK (8) / QUAD-FORK (color and smoothing added).

The final Indus sign discussed here has two rather different forms and might easily be described as two separate signs.  Originally I termed the first “variant” ASTERISK, noting in parentheses afterward that it contains eight points (VI32).  Tentatively, I maintain that name, calling this “A.”  Included in the same enumeration is the “B” variant, which I originally thought of as a DOUBLY DOTTED EX.  That is, the first looks like an “X” with another, curved “X” superimposed on it (“A” variant).  The second is an “X” with two short verticals rising from its upper “arms” and another two verticals descending from the lower “legs.”
The “A” variant appears elsewhere only in Wells’ list (W552), where it is given as a singleton from Harappa (H-43).  The “B” variant appears as KP245 and W556.  Wells gives it as another singleton (M-949).  I see the latter variant also on M-975, M-1205C, L-14, and L-48).  I cannot quite decide whether these two variants really are the same sign rendered somewhat differently.  It may be preferable to leave them as separate signs, as Wells encodes them.  In that case, they need better names.
Egyptian hieroglyphs include no exact parallel, although one which Gardiner calls “three fox skins tied together” is similar (F31).  It occurs in the word mst, “apron of foxes’ skins.”  The glyph resembles an asterisk with six points, the lower three longer than the upper three. 
In Old Chinese, two forms of one character are vaguely similar: mi3, “grains of different plants...separated by the thrashing” (Wieger 1965: 285).  One variation in the ancient writing is a “+” with diagonal strokes added between every two “arms.”  The other variation is an “X” with a short vertical stroke between every two “arms.”  The modern character for grain is based on the first variation.  Still in modern Chinese, the same character with an additional diagonal stroke at the top becomes bien4, “the steps of a wild beast....The examination of the trail..., hence...to discriminate, to part, to sort out” (1965: 286).
Proto-cuneiform contains an eight-stroke asterisk closely resembling our own symbol of that name: AN.  This came to designate “sky, heaven, the god An; grain ear, date spadix; to be high; high, tall; in front.”  As a prefixed element, it is also read DINGIR, “god/goddess,” acting as a determinative before names of deities.  An asterisk of six points, an “X” with a horizontal stroke through it, is HAL, a sign that came to have a great variety of meanings in later cuneiform.  Among them are “secret; divination expert; to stream, run; to divide, separate.”
Proto-Elamite also provides two types of “asterisk,” the eight-pointed type (M046) and the six-pointed type (M046~a).  There is also a sign with a resemblance to an asterisk formed of two “V” shapes crossing one another at an oblique angle (M102).  One variant has, in addition, four wedge-shaped impressions arrayed about the left half (M102~k).  The other variant of note has six smaller marks arrayed more symmetrically around it (M102~k2).
In rock art, as I noted in the discussion of the three-stroke ASTERISK (III19), there are many star-like forms.  In Texas, there are many examples with varying numbers of points, some odd numbers, including five and seven (note eight-pointed example Newcomb 1996: 53, Pl. 17, no. 6).  In the collection from Nevada and eastern California, the authors note 27 total occurrences of “stars” of three to six strokes (Heizer and Baumhoff 1984: 83; note the eight-pointed example p. 98, fig. 35g).  A star-like engraving also appears at more than one site in Australia: at Eucolo Creek, at the Rockholes, Panaramitee Station, and at Wharton Hill, Olary region (Flood 1984: 112, 183, 185). 
The closest motif to an asterisk that I find in African rock art is painted variously on the great overhang of Songo at Sanga, Mali (Le Quellec 2004: 60-61).  Some of these motifs seem at first glance to be six-pointed asterisks, most painted in red with a white surrounding line.  However, as the more detailed examination of six of these shows, they are not stars but representations of lizards (2004: 64, fig. 9). 
Among other African art forms, motifs more like the Indus ASTERISK appear.  For example, the Adinkra designs that are typically used on funerary cloths include a seven-armed, an eight-armed, and a nine-armed, asterisk-like motif (Willis 1998: 200-201).  One difference between these and the Indus sign is that the African examples surround a circle rather than a point (there is also a 10-armed asterisk that encloses a circled cross). 
Only the eight-armed example is pointed; the others generally have circles at the tips of the “arms” (a feature sometimes seen in American rock art also).  The star with eight arms is called nsoroma, “a child of the heavens, a star.”  There is a proverb associated with each of the Adinkra designs, that of the eight-pointed star being as follows: Oba Nyankonsoroma te Nyame so na onnte ne ho so, “A child of the Supreme Being, I do not depend on myself; my illumination is only a reflection of His” (1998: 154-155).  Thus, although the design depicts a star, it represents a person of excellent character or faith.  Variants of this symbol have other numbers of arms, as well: five, seven, or even ten.


Chinese: Keightley, David. 1978. Sources of Shang History: The Oracle-Bone Inscriptions of Bronze Age China. Berkeley: University of California.

Peruvian (Inca): Appleton, Le Roy. 1971. American Indian Design and Decoration. New York: Dover. (Originally published 1950 by Charles Scribner's Sons, entitled Indian Art of the Americas).

Monday, December 20, 2010

A Christmas Package: The Cross in the Indus Valley

Before I proceed with my discussion of six-stroke signs in the Indus script, I want to bring up a symbol commonly found on seals that is not a sign.  That is to say, it does not occur in the inscriptions but appears as a motif in various forms on seals that are mostly without inscriptions, or painted on the occasional pot shard.  This is the cross, a symbol best known to those in the West as a Christian emblem. 
P-16 and P-18: crosses or exes from
Pirak in the Indus Valley (color
and smoothing
by author).

As it happens, the cross appears widely around the world in contexts that are clearly pre-Christian or non-Christian.  These are the motifs that I will discuss in this post.  As an icon rather than a sign, the cross does not appear in any of the lists of Indus signs that I have cited (Koskenniemi and Parpola, Wells, Fairservis).  By my own count, then, there may be as many as 67 crosses in the corpus, the precise number depending on how one defines “cross.”  Distinguishing a cross from what I term an “X” is not a simple matter in this case, since the cross motif occurs alone.  One can view it from one angle or another and see either an “X” shape or a “cross,” as a result.  But, assuming that all of the possible crosses are that (and not “exes”), there are 11 from Mohenjo daro, 20 from Harappa, 30 from Pirak, three from Mehrgarh, two from Rahman-deri, and one from Nausharo.
Some of the seals from Pirak are themselves cross-shaped (Pk-3, -4, -5, and -8).  Others are round (Pk14, -15, -17, 22, -27, -29, and -33); still others are square (Pk-6, -19, and -10), and one rectangular (Pk-20).  In all of these cases, there are additional design elements: lines, dots (most often in the form of drilled holes), or both.  The occurrences from Mohenjo daro and Harappa are invariably quadrangular (M-349, -352, -464B through 466B, -1255, -1256, -1257, -1258, -1415B, -1416B; H-120, -121, -122, -171, 331E through -336E, -166B, -630, -634, -630 through -635, -637, and -638) .  Both of the examples from Rahman-deri are painted motifs on pot shards, one enclosed in a circle (Rhd-229, -233).  Of those cited previously, a series from Harappa may well have been intended as indicators that one side of a tablet was, in essence, blank, i.e., not inscribed (H331E through H336E). 
Cross (or ex) on square (or diamond-shaped) seal
from Altyn Depe (color added by author).

The archeological site of Altyn Depe in Turkmenistan has also produced a number of seals with crosses, dating to the Bronze Age (Masson 1988).  As in the Indus Valley, there are typically additional markings on these seals, including circles, lines, triangles, zigzags, and a few other elements.  There are examples of round, square, and rectangular seals (Pl. XVII nos. 6, 5, and 4, respectively).  But most are on seals that are themselves cross-shaped (Pl. XXI no. 7, Pl. XVI nos. 3, 5, 10, 14, and 15).  One seal is a crescent with a cross attached (Pl. XVII no. 15).  The presence of a seal here that bears two signs of the Indus script attests to contact with the contemporary civilization of the Indus Valley.  Further evidence of contact between Central Asia and South Asia during the Bronze Age is provided by seals described as having the shape of a “step-sided lozenge” appear in the Corpus of Indus Seals and Inscriptions, noted as being of “foreign” origin (Joshi and Parpola 1981: 365).  This "step-sided lozenge" shape (i.e., a cross-shaped seal elaborated by the superimposition of a central square) is quite common among the elaborated crosses of Altyn Depe.

Two elaborated cross shaped seals from Altyn Depe, the
"step-sided lozenge" and one with a central circle (color added by author).

Turning to other contemporary cultures, we note that Egyptian hieroglyphs include three symbols that are more or less cross shaped.  The simplest of these appears to be two planks, a tall vertical with a short horizontal attached below the central point.  This glyph may be ideographic in a word that translates as the English phrase, “who is in” (Z11).  The glyph also has a phonetic use based on the sound of the Egyptian word.  A second sign is less obvious, consisting of four simple loops arranged in a cross pattern (M42).  This may be a schematic representation of a flower, a glyph with the phonetic realization wn. 
Elaborated cross motifs on pottery from Altyn Depe (color and smoothing of image by author).  The "stepped" version at left appears on Hopi pottery in the American Southwest, where it is known as a squash blossom.

With the final Egyptian glyph, it is difficult to determine what object is represented (Aa27).  It comprises a vertical post topped by a small circle or oval.  On either side and on top of the circle there are small triangles, each attached by its apex.  For unknown reasons, this glyph has the phonetic form nd and is often combined with the small, round pot (W24, nw).  In hieratic writing, the distinction between this glyph and the previous flower is lost.  Both tend to become simple crosses.  In my earlier post on the Indus EX (II12) and that on the CIRCLED EX (IV42), we saw that there is also a CIRCLED FAT EX motif among Egyptian hieroglyphs (a circled containing four triangles that delineate an “X” shape in the center), an ideogrph in niwt, “village” and a determinative in town and village names and inhabited areas (O49).
In Chinese, both ancient and modern, a simple cross of two lines (of equal length) is shi2, “ten” (Wieger, 1967: 68).  Two verticals with a single long horizontal may be used for nien4 (or er4-shi2) “twenty” (1967: 70).  In the older script, one also finds three crosses in a horizontal row, the central one taller than the two on either side, as san1-shi2, “thirty” (1967: 71).  Today, “thirty” is normally written with two characters, as spoken, the three stacked horizontal strokes for san1, “three” preceding the cross shaped “ten.”  Note that a motif much like the Old Seal version of “thirty” sometimes appears in modern Christian art, where it represents Jesus on the cross at Golgotha, with the two thieves who were crucified at the same time on either side (reference below).

The American "stepped" version of a cross associated with Kokopelli, as photographed
from the wallpaper and bath curtain in a house in Texas (photo by author).

Turning to more distant cultural horizons, crosses appear in the art of the various native peoples of North America.  In the Southwest, two such instances occur in apparent composition with the motif of the flute player who is sometimes called Kokopelli (Malotki 2004: color plate 23 between pp. 48 and 49, near Holbrook Arizona; and Fig. 10e, p. 104).  In the first instance, the motif itself is a simple cross surrounded by a rounded outline of a cross (also found in South America at Hinkiori near Cuzco, Peru).  The second instance resembles the elaborated crosses on a square seal from the Indus Valley.  The cross shape itself is created by a motif of square coils with four-fold symmetry.
The specific type of elaboration of such crosses may be significant in distinguishing different meanings.  In modern Hopi culture, the sun appears as a circled cross, a circled “X,” and as a circle with four attached rays (the last of these three recalling the Celtic cross that is both Christian and pre-Christian).  On the other hand, a cross that is not encircled or one that has bent arms (i.e., a swastika) represents the earth to the Hopi (Waters 1963 and 1977: 93 and 114).  In contemporary sand paintings from the Navaho culture, in contrast, a circled cross may be either a basket with significant contents or a fire in a hearth, depending on context, as noted in a previous post.  More generally, elements of a great many Navaho sand paintings are laid out in the shape of a cross, with the arms oriented toward the cardinal points (Newcomb and Reichard 1975).

A swastika design on a seal from Altyn Depe, very similar to those found at
Mohenjo daro and Harappa (color and smoothing by author).

Before we move on to other examples of crosses, I return briefly to the subject of the swastika.  Although most people in the West now think of this as a Nazi symbol, it appears on many continents and in widely separated cultures, from a very period.  There are plenty of examples on Indus seals (e.g., just in the first volume of the Corpus: Mohenjo daro 332-347, 443, 435, and 448A; Harappa 104-118, possibly 119 depending on how one looks at it, 182B in a row of five, in the center of 242 with signs surrounding it; Lothal 29-72 and possibly 73).  The Chinese swastika is an early variant of fang1, about which Wieger states, “It is supposed to represent two boats lashed together..., a pontoon...by extension, square, regular, correct, a rule, etc.” (1967: 271).
A curved swastika appears in Africa on the so-called “fetish stone” on the Zaire River (Le Quellec 2004: 87, fig. 47, no. 30, on Pedro do Feiti├žo).  The African continent is home to an extraordinarily diverse group of peoples, many of whom incorporate a cross in their art.  In the north, at Tassili n’Ajjer in the Sahara, anthropomorphic and zoomorphic petroglyphs are sometimes accompanied by Tifinagh letters, including a cross (2004: 43, fig. 42).  In the intertropical zone south of the great desert, the boxed cross (cf. Indus WINDOW) occurs at Sanga, Mali (2004: 60-61, fig. 6).  A circled “X” occurs at the end of the arm of an anthropomorphic figure at Kourki, in Niger (2004: 58, fig. 3).  This figure may be interpretable as a feathered or horned warrior bearing a shield.  A simpler motif that closely resembles the Egyptian crossed planks appears on the great Sango overhang at Sanga (2004: 66, fig. 14).  This probably represents a spatula that local people use to churn milk.  Another boxed motif (WINDOW) appears at Ouoro-Kourou near Kita, Mali (2004: 71, fig. 22).  This is elaborated by the addition of two dots in each section, one red and one white.  A simple “X,” a “window,” and a Christian cross all appear at Rumfar Kurosha, Nigeria (2004: 77, fig. 30, nos. 2, 3, 4).
Still farther south, a doubled “window” and a cross made up two lines of dots occur in Kiantapo Cave in the Democratic Republic of Congo (2004: 84, fig. 44).  In neighboring Angola, there is a motif comprising a cross inside a doubled circle in Lunda country (2004: 89, fig. 50).  This occurs in rock engravings and in tattoos on the human body, where it represents the sun.  In the same area, another motif found in rock art and tattoos is almost identical to the Egyptian “flower” of four crossed loops. 
A golden circled cross from County Monaghan, Ireland (color and smoothing of image by author).

In Zambia, people trace elaborate looping motifs in sand (2004: 90, fig. 52 through 54).  The first of the cited motifs is not a cross, but is of interest because it is identical to an icon found on three Indus tablets (M-507B, M-508B, M-1457B).  In Zambia it is named liswa lyavandzili, “the nest of the ndzili birds.”  The second of the cited motifs is reminiscent of the interlacing cross ornamentation seen on some Pictish stones of Scotland (e.g., Fraser 2008: 32-33, fig. 35).  This type also resembles an interlaced star with five rounded “points” found on two Indus seals (C-49B and C-50B from Chanhujo daro).  The African symbol is found on the door of an enclosure used during circumcision rites, where it magically repels lion-men. 
The third Zambian design is an elaboration of this same motif.  In this example, there are two large loops, one crossing the other, with round indentations or dots arranged in and around various parts of the resulting cross.  This motif recalls the highly elaborated drawings termed rangoli or kolam in modern India (references below).  These are drawings on the floor or ground, made with powdered materials such as rice powder or crushed limestone, as part of the celebration of Diwali.  A new design is created each day of the festival to welcome Lakshmi.  Some websites explain these motifs as the footprint of the goddess or note that footprints are added near the drawing.  Many such elaborate designs showing four-fold symmetry are built on a cross-shaped central element.  Others are more like a star or daisy pattern, with many “arms” or “points.”  The center may also take the form of a swastika.  But in India, it is mostly women who make these designs, while in Zambia the artists are men.
East Africa also provides examples of a circled cross, especially in Kenya.  Two occurrences are found in Kakapeli shelter, Busia region, where rock paintings are used during initiation ceremonies (Le Quellec 2004: 130, fig. 42).  Cross motifs also occur among the list of signs engraved in cemeteries at Namoratunga in the northwest of Kenya (2004: 131, fig. 44).  In addition, such designs form cattle brands that are used by the Turkana people, representing various patrilineages.  The simple cross of two strokes is no. 57 in this list; the circled cross no. 7, and a circled cross of doubled lines no. 128. 

Proto-cuneiform token in form of circled cross of double lines (replica made by author).

The last of these recalls one of the circled crosses from Rahman deri in the Indus Valley (Rhd-233), as well as on very early protocuneiform seals from southern Iraq.  The latter symbol is known as SIG2~a3, which came to mean "hair, wool, fur."  This contrasts with the protocuneiform circled cross, UDU~a, "sheep; small cattle."
We can even find the circled cross in Australia (in a painting at Uluru/Ayers Rock, Northern Territory) and on the island of Tasmania at Greens Creek (Flood 1997: 156 and 232).  These must not be representations of spoked wheels, in most cases, since the native peoples of these areas did not use such a wheel before Europeans arrived.  But in modern times, at least one occurrence of the circled cross does represent a wheel.  It appears in a representation of an automobile (1997: 216, at the Granites, Northern Territory).
Thus, the cross appears on every continent, its first occurrences long before Christianity.  Such distribution is unlikely to be due to borrowing.  As it is either a very simple motif or an elaboration of one, I think it more likely that widely divergent social groups devised it independently and assigned it meaning.  Thus, it would be foolish to look at just one of the meanings associated with a cross and apply it uncritically to the cross shapes of the Indus Valley.

Le Quellec, Jean-Loic. 2004. Rock Art in Africa: Mythology and Legend.  Transl. Paul Bahn. Paris: Flammarion.

Malotki, Ekkehart. 2000. Kokopelli: The Making of an Icon. Lincoln: University of Nebraska.
Waters, Frank. 1963 and 1977. Book of the Hopi. Drawings by Oswald Bear Fredericks. New York: Penguin.

Indian rangoli and kolam:
Wieger, Dr. L. 1965. Chinese Characters, Their Origin, Etymology, History, Classification and Signification. Transl. by L. Davrout. New York: Paragon and Dover.  (Orig. pub. by the Catholic Mission in 1915).

Fraser, Iain, ed. 2008. The Pictish Symbol Stones of Scotland. Edinburgh: Royal Commission on the Ancient and Historical Monuments of Scotland.
and Christian:
The Pious Child: Instructions and Prayers for Catholic Children. 1894. New York: Benziger Brothers, Printers to the Holy Apostolic See.

Thursday, December 16, 2010

An Indus Rake and Comb

This post discusses four signs in the Indus script, each comprised of six strokes.  The first of these contains two tall vertical lines in the center.  On the right of these and on the left are two short vertical strokes, one over the other.  Thus, the two posts in the center are surrounded by four short strokes, similar to the keyboard sequence : | | : although not identical.  Wells does not list this particular sign, but he uses the term “caged” as a description of this type of modification of a basic symbol by surrounding it with four short strokes.  Following his example, I call this sign CAGED DOUBLE POSTS and enumerate it V17.  It appears elsewhere only in the list of Koskenniemi and Parpola (KP146). 
Inscription M-880: STRIPE BELTED AITCH / CUPPED POST / POTTED ONE / FAT EX IN DIAMOND / VEE IN DIAMOND / POT  (detail, over head of unicorn bull; smoothing of image and false color added by author).

Sign V17 does not appear in my own database of inscriptions any more than in the sign lists of Wells and of Fairservis.  The KP sign may have been derived by inference from M-1201F, an inscription on the short side (not the face) of a broken seal.  To the right of the break in the upper part of the seal's face (looking at the side edge) there is an abraded post, then a clearly preserved post, hence DOUBLE POSTS.  This is followed by the two short strokes stacked one over the other that is typical of caging.  In my database, I coded this DOUBLE POSTS / STACKED TWO, because that was all that I saw.  Apparently, Koskenniemi and Parpola proceeded from such an observation to the inference that originally there must have been another “stacked two” on the other side of the DOUBLE POSTS.  This is quite possible.  However, since the first “post” is marred by the break, we cannot tell whether there was also another post (or more, or even another symbol) before this hypothetical, reconstructed “stacked two.”  
In addition, the STACKED TWO occurs at least once by itself, not part of a “caging” pair.  Thus, it is also possible that there was no other symbol on the other side of the DOUBLE POSTS.  In any case, this seems to be the only occurrence of this combination.  On a different seal from the same city, M-137, there is a sequence "stacked two" plus CARTWHEEL / DOUBLE POSTS and another "stacked two."  Thus, I term this whole sequence CAGED CARTWHEEL & DOUBLE POSTS.  It is certainly possible that the unbroken inscription on M-1201F was originally a duplicate of that on M-137.
Although it is a very simple combination of elements, I have not observed this exact symbol (CAGED DOUBLE POSTS) elsewhere.  Among Egyptian hieroglyphs, there are two that include four dots or short strokes, U9 and U10.  In both of these, the main part of the sign resembles a barrel turned on its side.  Extending from one end come four dots in a diagonal arrangement, seemingly pouring out of the overturned barrel.  This is the form of U9, representing a grain measure with grain pouring out.  This glyph occurs at the end of the word “grain” as a determinative.  In U10, the same symbol also has three small diagonal strokes above the “barrel” section.  This is a combination of glyph M33 (three grains above) plus U9, which together serve as an ideograph in “barley” and/or determinative “emmer (wheat).”  Thus, four short strokes do appear in Egyptian as part of a glyph, but not “caging” another glyph.
In proto-cuneiform, one sign is made up of two long horizontal strokes above and below two shorter strokes (ZATU772).  This is essentially the reverse of the Indus sign, using long strokes to bracket short ones.  A similar situation occurs in proto-Elamite, where two long horizontal strokes form one sign (M009).  A second sign, which may be a variant of the first, has these same long strokes above and below a set of short strokes (M009~h).  In this case, there are five internal shorts, two in a stack on the left, three in a stack on the right.  Once again, the double posts of the Indus sign are replicated.  But these are not bracketed by the stacked short strokes.  Instead, the long strokes bracket the short ones.
The second sign considered here is the STRIPE BELTED AITCH (VI18, the eighteenth of the six-stroke signs).  It resembles our capital letter “H” but with two horizontals instead of one, and there are additional short strokes between the two horizontals, joining them.  This sign does not appear in Fairservis but is KP298(b) and W185.  Wells shows it with three short strokes between the horizontals while the KP form has just two.  Wells notes VI18 to be a singleton, appearing only on M-880.  The Corpus indeed shows three short strokes between the two “belts,” revealing VI18 to be a seven-stroke sign in reality.
There are two hieroglyphs in the Egyptian list with some resemblance to the Indus sign.  The first of these represents a coffin (Q6).  It is wide and low, with no lines joining the two central horizontals.  In addition, over the top horizontal is a curved line, indicating the curved top of the coffin.  Since this glyph is a representation of a coffin, the horizontal elements are wide apart, a feature not seen in the Indus sign.  The second Egyptian glyph represents the mast of a ship (P6).  This is a tall and thin “H” shape (more or less) with four crossing horizontals, perched on top of a “Y” shape.  Again, the similarity to the Indus sign is remote.
Old Chinese provides a better parallel with dan1, “cinnabar....The crucible or stove of the alchimists [sic], with cinnabar in it” (Wieger 1965: 270).  The basic shape is again much like an “H,” and there are two horizontal strokes.  However, instead of two or three short verticals joining the two horizontals, there is a single dot between them (the cinnabar in the crucible).
In proto-cuneiform, two signs somewhat resemble Indus VI18, but rotated 90 degrees.  There is ISZ~a, two horizontals joined by two verticals, with a shorter horizontal to the right of the right-most vertical and another to the left of the left-most vertical.  It seems to be the letter "H" fallen on its side, with an additional post down the middle, the center of which is hidden by the belt formed by the crossing strokes.  The second sign, U2~c, is identical except that the shorter horizontals in the center join, making a single long horizontal.  The first symbol came to mean “mountain,” and the second, “plant.”
Proto-Elamite contains three signs with some resemblance to this.  There is an “H” shape on its side with two “belts” (M026).  A possible variant of this adds an “X” between the two “belts” (M026~h).  A third moves the two “belts” to the left end, adding three short strokes to join them (M032~b).

Inscription M-1295: (POTTED ?) / CRAB / RAKE ("a") / SLASHES IN OVERLAPPING CIRCLES / QUINT-FORK (bar seal, with smoothing and false color added by author).

Our third Indus sign, VI19, has been called a “rake” elsewhere and I repeat this term.  The RAKE appears elsewhere as KP92, W264, and Fs Q-11.  Fairservis relates it to the COMB, which it resembles, turned horizontally and given the affix P-1 (i.e., the SINGLE QUOTE, which he previously defined as the genitive/possessive).  The COMB, according to this author, means “write, scratch; dative of person, belongs to; women’s cloth.”  Nevertheless, when put together, a horizontal COMB on top of a SINGLE QUOTE (which looks more like a SINGLE POST to me), the meaning becomes “of below, downriver, south.”
Wells notes 70 occurrences of VI19 in three variants.  There are 54 occurrences from Mohenjo daro, 10 from Harappa, three from Lothal, and one apiece from Kalibangan, Pabumath, and Rupar.  Wells distinguishes the versions based on the number of prongs on top, his "a" having five of these, "b" four, and "c" six.  I find more than three variants, five based on the number of prongs (4-8), and possibly an additional variant to account for those of a smaller size.  If we ignore the size distinction for the moment, we can still call the most common variant with five prongs “a,” the second most common variant with four prongs “b,” a relatively rare variant with six prongs “c,” a very rare variant with seven prongs “d,” and a singleton variant with eight prongs “e.” 
Mohenjo daro provides examples of all five variants (rare variants: M-163 the “c” type, M-279 the “d” type, and M-406 the “e” type).  The last two are smaller than the other signs in the inscription, fitting partially or completely underneath BI-QUOTES, a sign I type to be intended as preceding.  Harappa yields “a” variants mostly on seals, “b” variants mostly on tablets, and one instance of a “c” variant (H-103).  All three instances from Lothal are of the “a” variant while the single occurrence from Kalibangan is “b” (L-18, L-98, L-145; K-54).  Those from Pabumath and Rupar are the “b” as well, and I add to these an instance from Banawali, also “b” (Pbm-1, Rpr-1, B-11).
Proto-cuneiform provides two partial parallels to the RAKE.  GAL is a horizontally positioned “rake” with six long prongs and GAL~b a three-pronged variant.  This sign came to have various meanings including “big, great.”  The six-pronged variety appears as part of the word SZANDANA, which came to mean “gardener.”  Proto-Elamite has three rake-like symbols with four, five, and seven prongs (M038~a, M0138~b, M038~d).  All are horizontal like the proto-cuneiform type, a feature unlike the Indus sign.

Inscription M-1390A on round tablet, reading from right to left: STOOL / PINWHEEL / COMB (false color and smoothing added by author).

The COMB with five “teeth” is sign VI20 in my list (twentieth of the six-stroke signs).  Printed with six “teeth,” it appears as KP98.  Shown with five “teeth,” it is W287 (prongs on the left side of the vertical stroke) and W282 (prongs on the right side of the vertical).  It is also Fs L-9, where it is said to represent a comb (hence my term for it).  Fairservis gives his suggestion for the meaning as “write ‘this mark’; scratch; dative of person, as terminal in texts ‘belongs to’; women’s cloth.”  These disparate definitions are lumped together as being semi-homophonous in Dravidian languages.
Wells divides the COMB into two different signs based on the direction of the “teeth.”  In his listing, W287 has the prongs on the left side of the vertical, in a singleton from Kalibangan (K-77).  The other sign, W282, is quite common, occurring 143 times in all.  Mohenjo daro provides 40 of these, Harappa 97, Kalibangan three, and one apiece from Chanhujo daro and Nindowari damb.  Wells divides these among six four-, five-, and six-pronged variants of two prong types, horizontal and diagonal. 
Unusual bird motif on M-1390B, the back of the round tablet.  This resembles bird motifs found in the art of Central Asia, e.g., Altyn Depe in Turkmenistan).

In my own database, I did not distinguish by prong type originally.  I had the impression that the diagonal forms occur only on tablets while the horizontal forms occur on both seals and tablets.  Thus, I considered the distinction to be based more on the style of signs on the whole artifact.  I did distinguish among the numbers of “teeth.”  In going back over these, I find my general “rule” to be wrong.  At Mohenjo daro, both seals and tablets show diagonal and horizontal prongs, numbering four, five, six, seven, and eight.  Those at Harappa also show a highly mixed pattern of diagonal- and straight-pronged variants, numbering three, four, five, six, seven, and eight. 
But there are also more variants just counting the number of "teeth" in the comb.  From Kalibangan come both diagonal and straight types again, the prongs numbering five, six, seven, and thirteen.  Lothal provides only two instances, one with four long horizontal prongs (L-96) and a more enigmatic version with two or three (L-97), both rather crudely executed.  Chanhujo daro yields three instances by my count rather than just one (Wells' count).  One has five diagonal prongs (C-14) and two have horizontal, numbering five and six (C-15, C-32).  There may be one from Rahman dheri with seven prongs (Rhd-267), a very faint sign on a pot shard (which Wells did not mention).  One is on a seal from Nindowari damb with six horizontal prongs (Nd-2), another on a seal from Rupar with five (Rpr-1).
Besides seeing more variety in prong number than Wells, I note the presence of instances with longer prongs than typical (M-874, M-1305, M-1311, H-226, H-227, H-228, and L-96).  The greatest difference between our observations is in the distinction between the COMB with prongs on the right side of the vertical and the REVERSED COMB with prongs on the left.  It seems to me that generally prongs point toward the beginning of the inscription.  Since many tablets show a sign order that is the reverse of that on most seals, a great many instances of the COMB (in my database) are in the reversed position and this is particularly common on tablets.  If all instances can be explained in this way, then it would be best to consider the COMB and REVERSED COMB as variants of a single sign rather than two different signs.
Korvink includes the COMB among the signs he calls "terminals" (2007: 28-34).  In discussing the "harrow," as he terms it, he notes that VI20 occurs both before and after two other terminal signs in some inscriptions (before or after the SPEAR and the POT).  The COMB follows one of these other terminals 232 times, preceding one of them only 19 times.  He does not attempt to explain the significant minority of occurrences preceding other terminals, only stating that this cannot be explained by “scribal error.” 
It is not clear to me whether or not his analysis accounts for the reversal of inscriptions on tablets, i.e., the fact that sign order is from right to left on typical tablets but from left to right on most seals.  Despite this generally tendency, I note that a few seals appear to reverse the usual sequence and read from right to left like tablets.  At the same time, some tablets appear to reverse their usual sequence and read from left to right.  If one assumes that the appearance of a common prefix (such as CIRCLED VEE / BI-QUOTES) indicates the beginning and the common terminal indicates the end of an inscription, many such apparent anomalies can be accounted for.  That is, one can simply choose to read a given inscription by starting with the prefix -- whichever side it occurs on -- and procede toward the terminal. 
However, this is circular reasoning, interpreting certain signs as prefixes and others as terminals and, in effect, insisting on a consistency in the corpus which will suit the hypothesis.  We need some other evidence one way or the other, before confidently assigning reading direction.
As far as parallels to the COMB go, Egyptian hieroglyphs do not provide much to choose from.  There is only a rough similarity between the Indus COMB and the harpoon-head (T19).  The glyph is comprised of a tall vertical from which two short horizontals extend near the top and two diagonals near the bottom.  The higher of the diagonals ends in a coil.  The Old Kingdom form of the same glyph has a single upper horizontal and neither of the two lower diagonals is coiled (T20).
Luwian hieroglyphs include a slightly curved, comb-like sign for the syllable mu.  This may be a schematic bovine horn since another symbol for the same syllable represents a bovine head.  Then again, these could simply be two unrelated ways of writing a single syllable.  It is interesting that mu is what English speakers tell their toddlers the cow says, spelling it "moo."  In linguistics, this "word" is termed echoic because it echoes a sound in the environment.
Proto-cuneiform has a much better parallel, a comb sign with nine long “teeth” (ZATU753).  In addition to this difference in the number of prongs, it has a horizontal line extending from between the 4th and 5th “teeth” on the left out beyond the rest of the sign on the right.  This additional line has no counterpart in the Indus sign.  Proto-Elamite also contains a few comb-like elements, all of which have their prongs descending from a horizontal line.  One has three prongs, another four, and a third variety five (M041, M041~c, M041~d respectively).  The meanings of these and of the previous proto-cuneiform sign are unknown.
Old Chinese does not have a precise parallel, at least not among the forms discussed by Wieger.  But among the inscriptions shown in his appendix, one variant of a character does appear, resembling a horizontal comb with prongs upward (Wieger 1965: 374).  This is a representation of mountains, a variant of shan1.  The character is now considerably more comb-like, written with a horizontal stroke and three verticals.  The ancient variant shown in this particular inscription has a the slightly curved base typical of the ancient form and five “teeth” (an atypical number of mountain peaks).
Motifs resembling the COMB commonly occur in the rock art of North America while the RAKE typically does not.  There are 306 occurrences of a “comb” in Nevada spread fairly evenly across all regions of the state where petroglyphs occur (Heizer and Baumhoff 1984: 83-84).  These are most often horizontally positioned, with prongs hanging down.  A few, though, have upward-pointing prongs.  Others have a diagonal base. 
There is also an element similar to the “comb” but with an additional line attached to the horizontal at one end and curving back over it.  This may be rain symbol; it occurs just twice and only in southeastern Nevada.  Instead of symbolizing rain, it may be a schematic representation of a mountain sheep, an animal otherwise commonly depicted in a less angular manner.  A third motif is something the authors term a “convoluted rake.”  This has the same appearance as the basic “comb” but one (or more) of the descending “teeth” curves back on itself.  Such an element appears 11 times, mostly in western Nevada and in one location in the east. 
Examples of “combs” with both upward and downward prongs are easily observed in this collection (1984: 126, fig. 63e and f).  The motif with upward prongs has seven of these while there are six downward prongs in the second example.  A third example on the same page has nine “teeth” protruding from a diagonal long line (1984: 126, fig. 63d).  The possible “rain” motif appears  nearby (1984: 128, fig. 65h).  It has five downward “teeth” and an upper curve that is indeed quite like representations of the horn on a sheep.  The fifth “leg” may indicate the animal’s tail or the male sexual organ.  A possible “convoluted rake” occurs at the Mouse’s Tank site (1984: 136, fig. 73e).  Five “teeth” descend from a horizontal in this case, with two prongs joining a circle below on the left and three joining a circle on the right.
Similar motifs appear in the rock art of Texas.  One panel alone contains four different “combs,” two with five “teeth,” one with six, and one with perhaps ten (Newcomb 1996: 74, Pl. 37, no. 1).  Historical tribes in this area participated in a mescal bean cult that may be referred to for explanations of many elements in rock art.  For example, Wichita and Pawnee peoples held initiation ceremonies for medicine societies during which new members were tested by scraping toothed jaws of garfish over their bodies (1996: 75).  It may follow from this that the "combs" here are representations of the ritually significant garfish jaw. 
However, as the author notes, there is a huge gap in time between the date of the petroglyphs and that of the modern tribes.  Connecting the two is a hypothesis, not a proven fact.  In addition, other “combs” may not fit this hypothesis.  One instance, for example, has over 50 “teeth” (1996: 85, Pl. 46, no. 4).  A single panel at Bee Cave has a horizontal, curved version with four upward prongs, two straight vertical versions with five “teeth” each, and a diagonal version with six “teeth” (1996: 97, Pl. 55, no. 1).  In another instance, a possible “rain” element occurs with four upward prongs, a short curve at the left end of the long stroke, plus two prongs extend beyond the others and then join. Hence, this may be a “rain” motif plus a “convoluted rake” combined.  This much variation (especially in one relatively small area) would be rather astonishing for a single symbol.
Before leaving the COMB, I will note in passing the ligatures in which it appears in the Indus script.  There is a COMB ON BATTERY, discussed previously in connection with the TRI-FORK TOPPED BATTERY (M-649), as well as a MAN HOLDING COMB symbol which appears at the same city (M-16, M-494, M-704).  The CIRCLED DIAMOND AND COMB occurs at Mohenjo daro, while a similar CIRCLED BISECTED RECTANGLE AND COMB occurs at Harappa (M-481, H-558).  The rather odd AY ON COMB seems to have a horizontal comb under one “leg” (M-357, M-470, H-158, H-189, H-190, H-208).  Possibly related, the BI-FORK ON COMB also has a horizontal comb as its base element in a seal impression from Desalpur (Dlp-3).  Kalibangan yields an example of CIRCLE WITH ATTACHED COMB (K-104) and CIRCLE-HEADED BEARER WITH ATTACHED COMB (K-79).  Another ligature with a horizontal comb, this time across the middle, is from Mohenjo daro: COMB BELTED ASTERISK (M-391).  The same is true of COMB BELTED MAN (M-142, M-488, M-831, M-1160, M-1162, M-1611, M-1612).
Inscription on tablet H-897A, from left to right: STRIPED DOOR WITH KNOB / BI-FORK TOPPED POT / POT / COMB ("a2") (smoothed and false color added by author).

*Note: two bird-shaped "seals" were found at Altyn Depe, one in excavation 7, priest's tomb, room 10, burial 235 (Plate XVII, no. 9) and another in excavation 9 from 1981, locality 254 (Plate XXXVII, no. 4).  Both are made of metal, the second more specifically bronze (Masson, V.M. 1988. Altyn-Depe. Transl. Henry N. Michael. Philadelphia: The University Museum, University of Pennsylvania).