Tuesday, February 21, 2012

Indus Valley Civilization and Cross-Cultural Contacts


Proto-Elamite bucranium or bovine head, a motif found on pottery (Potts 1999: 53)


During the third millennium BCE (between 3000 and 2000 BCE), early civilizations made contact in an area termed “Middle Asia” by the archeologist G. Possehl (2002: 215).  The Indus Civilization was the easternmost part of this area.  It also included the southwestern portion of Central Asia, which Possehl refers to as Turan; southeastern Iran and the western border of the Persian (or Arabian) Gulf, known to Mesopotamians as Magan or Makan, as well as islands in the Gulf such as Bahrain; and Mesopotamia itself, the land between the rivers Tigris and Euphrates.  Contemporary kingdoms to the west in Egypt and to the northwest in Canaan were peripheral to “Middle Asia,” although also in contact.

Gulf seal with bucranium (top center), anthropomorph (left), grid,
and scorpion (right), as well as bird (Kjaerum 1983: 37).
There is an Intercultural Style of objects carved in soft stones (including steatite, the material of most Indus seals), common to this area.  Certain motifs appear in this style in several locations.  There are scenes of combat with snakes, depictions of bulls including the zebu or humped bull (seen on some Indus seals), a lion-headed bird called Imdugud in Mesopotamia, depictions of huts or temples resembling the Indus VEST sign, plant-like forms including rosettes and palm trees, and geometric patterns such as grids, whirls, guilloche, and an imbricate motif (nested arcs) (Possehl 2002: 216-7).  The shared motifs may indicate a shared ideology of some sort.  Certainly, though, these are indications of contact and trade, much was which was apparently conducted via waterways.

An Indus bucranium on a pot shard (Shah and Parpola 1991: 374, Rhd-241).

During the Early Dynastic Period in Iraq, the lands of Magan and Meluhha occasionally appear in cuneiform literature.  The early Akkadian king, Sargon, ruled at the time when these lands were known (Possehl 2002: 218).  Copper was mined in Magan during this period and the same metal is mentioned in cuneiform documents as coming from Meluhha (2002: 220).  Other products from Meluhha include semi-precious stones such as carnelian (a red stone), lapis lazuli (a blue stone), and pearls.  A few wood products, a few animals, fresh dates, and some gold also came from Meluhha.  Many more products derive from Magan, including ivory, diorite (a hard, volcanic stone), semi-precious stones, red ochre, special types of wood or objects made of wood, goats, and gold dust (loc. cit.).  The Gulf was a center of this international maritime trade.

A zebu on a plaque from the Elamite Diyala Valley (Lamberg-Karlovsky and Potts 2001: 225).

Four depictions of boats are known from the Indus Valley, three from Mohenjo daro, one from Lothal (2002: 221).  There are many other boat depictions known from Mesopotamia, as well as some from the Gulf.  One witness to this trade comes in the form of stamp seals.  According to Possehl, over two dozen seals have been found in the Near East that either come from the Indus Valley or resemble Indus seals (2002: 221).  They date from the late third millennium to the Babylonian Kassite Period.  Some bear pictorial motifs that resemble those on Indus seals, while others have Indus script or symbols resembling those of the Indus.  Beads are another witness to inter-regional trade, especially beads of etched carnelian.  These were found in considerable quantities at Mohenjo daro, Harappa, Chanhujo daro, Kalibangan, Lothal, and Rojdi in the Indus Valley.  They also appear in Mesopotamia at Ur, Kish, Tell Asmar, Tell Abu Salabikh, and Nippur.  Others show up in Elamite cultural areas including Tepe Hissar, Shah Tepe, Susa, Tepe Yahya, and other locations.

Seal M-1111 from Mohenjo daro with inscription: FAT EX / STRIPED FAT CEE /
POTTED THREE / CRAB / POT (note how different the zebu is from the Elamite example).


Dice are another intercultural phenomenon, both in the cubical form known to the modern world, and “stick dice” (2002: 224).  Various objects of other types also occur in the Indus Valley, Mesopotamia, and the Gulf: some types of pottery, a few figurines, spiral-headed pins of copper or bronze, , and Harappan weights.  Oddly enough, though, very little material from the west has been found in the Indus Valley (2002: 227).  Possehl notes six Gulf seals and six cylinder seals, a few metal objects, and the motif formerly described as Gilgamesh battling two animals.  In Mesopotamia, this scene typically shows a hero with a lion, a bull, or between two such animals.  The comparable Indus scene contains two tigers instead.
Panel from an inlaid harp from Ur, Mesopotamia, showing
"Gilgamesh" with two bull-men (Aruz 2003: 106).

Seal M-306 with hero between tigers, the "Gilgamesh" motif
(inscription VEE IN DIAMOND / DOWN MAN ON BASE / POTTED ONE /
CIRCLED FORK / CRAB / POT / TRIPLE BRICK).

Possehl notes one type of artifact that appears to derive from as far afield as the Minoan Crete culture (2002: 228).  The object is a ceramic ring with small cup-like elements attached, a ring-kernos.  The ring is hollow, the cups perforated at the bottoms so that liquid poured into the cups would also fill the ring.  This specialized vessel was later used in the Kernophoria, a Greek festival of the harvest.  Another archeologist, J.M. Kenoyer, cites the pictorial motif of bull leaping as well (2003: 408).  This peculiar activity is familiar to Westerners as characteristic of Minoan Crete.  But there are also at least two Indus seals that seem to bear the same theme (see M-312).  Kenoyer mentions another from Central Asia (2003: 409).

Seal M-312, possibly showing bull leaping.


The Indus Civilization also had contact with a Central Asian area generally termed Bactria Margiana, centered around the Amu Darya river.  This area and parts of eastern Iran form Possehl’s Turan.  The semi-precious stone lapis lazuli occurs in this area, which may account for the presence of Harappan artifacts.  But the earliest evidence of contact is the appearance of pottery with a stepped cross motif, a pattern later found both in Central Asia and in the Indus Valley.  From the Mature Harappan period, there is a stamp seal with two Indus signs from Altyn Depe in Turkmenia.  Another stamp seal is clearly of Indus origin, bearing both signs and the motif of the elephant, this object coming from Gonur Depe.  Moving in the opposite direction, one seal from Harappa bears an eagle that closely resembles seals from Gonur Depe (2002: 230).  The other side of the Harappan seal shows the stepped cross already cited.
Seal H-166A showing a bird (eagle?) with spread wings.
Elamite bird (eagle?) with spread wings on an axe head from
Tepe Yahya (Lamberg-Karlovsky and Potts 2001: 216).
Thus, although many of the details of the trade among the Indus Valley, the Gulf, Mesopotamia, and outlying regions remain unclear, the Harappans had far-reaching contacts during the Bronze Age.

REFERENCES

Aruz, Joan, ed. 2003. Art of the First Cities: The Third Millennium BC from the Mediterranean to the Indus. New York: Metropolitan Museum of Art and Yale University.

Kenoyer, J.M. 2003. "The Indus Civilization," in Art of the First Cities, Joan Aruz, ed. New York: Metropolitan Museum of Art and Yale University. Pp. 377-380.

Kjaerum, Poul. 1983. Failaka / Dilmun, the Second Millennium Settlements, Vol. 1:1, The Stamp and Cylinder Seals. Jutland Archaeological Society Publications XVII:1. Moesgard, Aarhus: Jysk Arkaeologisk Selskab.

Lamberg-Karlovsky, C.C. and D.T. Potts. 2001. Excavations at Tepe Yahya, Iran, 1967-1975: The Third Millennium. Cambridge: Peabody Museum of Archaeology and Ethnology, Harvard University.

Possehl, G.L. 2002. The Indus Civilization: A Contemporary Perspective.  Lanham, MD: AltaMira.

Potts, D.T. 1999. The Archaeology of Elam: Formation and Transformation of an Ancient Iranian State. Cambridge: Harvard University.

Shah, Sayid Ghulam Mustafa Shah and Asko Parpola. 1991. Corpus of Indus Seals and Inscriptions. 2. Collections in Pakistan. Helsinki: Suomalainen Tiedeakatemia.

Thursday, February 9, 2012

A Study on Segmenting Indus Texts

Detail from seal M-174 with inscription showing Yadav et al's four-sign "beginner"
CARTWHEEL / BI-QUOTES / TWO POSTS / FISH (plus DOUBLE CEES / POT)

I recently finished reading the summary of Indus script research by archeologist G. Possehl (1996).  None of the proposed decipherments receives full support from this author, an accurate statement of the progress (or lack thereof) in this field.  Possehl notes the obstacles to further progress, including the lack of agreement on a basic sign list which is an important first step!  He also gives a short list of characteristics of the script generally agreed upon: reading direction generally from right to left; the presence of suffixes but neither prefixes nor infixes; the number of signs (whatever the accurate count is) indicates a logographic script rather than alphabetic or logographic. 
Even this short list is arguable, as noted in various posts.  Seals appear to read left to right (the POT occurs on the right), which becomes the proposed right to left in impressions (the POT occurs on the left).  Most objects termed tablets would not make good impressions if used as seals and appear to read right to left.  There are a few apparent exceptions, however, which is not surprising for the Bronze Age.  As Korvink's analysis shows, there are terminal signs, but the interpretation of these as suffixes is unwarranted at this stage.  If this script does not provide phonetic information -- if it is not writing but is, instead, proto-writing -- then the terminals are almost certainly not suffixes.  The prefixes and infixes supposedly absent might actually have been present in the speech of the ancient Harappans, just not written.  The difference is that between writing and language.  It is best at this early stage of progress to separate symbol and speech, to analyze the script as a semiotic system without drawing linguistic conclusions.
Detail from seal H-444 showing Yadav et al's four-sign "beginner"
VEE IN DIAMOND / BI-QUOTES / 2 POSTS / FISH (plus CAGED WHISKERED FISH).

I have often referred to M. Korvink's study, the first I had seen that clearly separated statistical analysis from linguistic issues.  Another is that by N. Yadav and colleagues, published too recently to be included in Possehl’s summary of work on the script (1996).  Yadav et al completed a study on sign sequences that I will review here (2008: 53-72).  Focusing on a select group of inscriptions from Mahadevan’s concordance, they seek out signs that typically come at the beginning, those that come at the end, and groups that tend to occur together in the middle of inscriptions (2008: 55).  In a sense, this group of researchers is looking for the same sorts of positional regularities as those Korvink examined (2008).  But because they go through their analytical steps in a different sequence, they arrive at somewhat different conclusions.  For example, they provide "Table 3: Beginner Four-sign Combinations," from which I excerpt the following (2008: 57):

Beginning Four-Sign Combination
Frequency
CARTWHEEL / BI-QUOTES / TWO POSTS / FISH
4
VEE IN DIAMOND / BI-QUOTES / TWO POSTS / FISH
9


In both the inscription segments in this table, Korvink’s analysis indicates that the first two signs comprise one entity (which he calls a prefix), while the last two signs are a frequent pair.  The two segments – prefix and medial segment (or PM) – only come together as an apparent four-sign combination because they include two of the most common prefixes and a common pair.

The same problem exists with Table 5: "Ender Four-sign Combinations" (2008: 58):

Ending Four-Sign Combination
Frequency
SINGLE QUOTE / AY ON QUOTES / CIRCLED FORK / POT HATTED BEARER
5
BI-QUOTES / RAKE / WINGED MAN / POT
4
FISH / MALLET / FORK / POT
4
POTTED ONE / MALLET / FORK / POT
4
FISH / CUPPED SPOON / 3 POSTS / SPEAR
4
DOT IN FISH / CUPPED SPOON / 3 POSTS / SPEAR
4
BI-QUOTES / CUPPED SPOON / 3 POSTS / SPEAR
4
MARKED FISH / CUPPED SPOON / 3 POSTS / SPEAR
5
CORN HOLDER / TRIPLE TRIANGLES / CRAB / POT
5
CUPPED SPOON / 3 POSTS / FOOTED STOOL / PINWHEEL
4
FISH UNDER CHEVRON / WHISKERED FISH / POT / MAN
4
PRAWN / ZEE / CROSSROADS EX / POT
16
MARKED FISH / BED / BOAT / POT
4
WHISKERED FISH / RAKE / WINGED MAN / POT
4



In all these examples, Korvink’s analysis demonstrates that the last sign is a terminal.  In one or two cases, the last two signs together comprise the terminal (POT + MAN is a two-sign terminal and Korvink treats FOOTED STOOL + PINWHEEL as another of this type).  In three of the examples, the constant sign that ends the prefix has been included in the purported four-sign group (SINGLE QUOTE in the first example and BI-QUOTES in two subsequent ones).  Thus, none of these is clearly a four-sign group. 
Seal H-61 showing Yadav et al's four-sign "ender" PRAWN / ZEE / CROSSROADS EX / POT
(preceded by FIGURE EIGHT WITH LADDER / PINCH / SINGLE QUOTE).

The best example is the one with the highest frequency: PRAWN / ZEE / CROSSROADS EX / POT with 16 occurrences.  The first three signs probably do form a combination, but POT is simply the most common terminal.  The Koskenniemi and Parpola concordance shows PRAWN occurring alongside ZEE but without the following CROSSROADS EX 18 times; all three occur together a total of 27 times (1982: 52-54).  (There are more examples in the KP concordance than shown in Yadav et al's table because the latter delete all broken or partially illegible inscriptions from the sample they study.)  In two KP inscriptions, a sign other than a terminal follows the three-sign combination (H-20 has HAIRY HUNCHBACK, H-278-284 have MALLET + FORK) and in two others, the sign following the combination is illegible (H-579, KP 6058).  Thus, since the combination can occur with or without the terminal immediately afterward, the terminal is not properly considered part of the combination itself.

The sequence containing CRAB brings up an interesting feature of that sign’s occurrences (Koskenniemi and Parpola 1982: 54-55).  This symbol often appears at the beginning of the medial segment (45 inscriptions).  On occasion, in this position it forms the whole of the medial segment, alongside a prefix or terminal (or both).  But it can also occur in the prefix, in which case it stands just before the prefix constant (5 or 6 inscriptions, depending on whether the similar symbol with multiple “legs” is a variant).  Then, its other most common appearance is as the final sign in the medial segment following other medial signs and preceding a terminal (37 inscriptions).  It is in this last group that the sequence CORN HOLDER / TRIPLE TRIANGLES / CRAB occurs (preceding the terminal POT HATTED BEARER in M-379; preceding the terminal pair FOOTED STOOL + PINWHEEL in KP 1588; preceding POT in H-191-193, H-728-731, M-51, L-12, M-92, KP 7132A, H-515, M-892, and M-665).  In one inscription, the same three-sign combination occurs before additional medial signs (M-471).  Thus, I see CORN HOLDER + TRIPLE TRIANGLES + CRAB as a combination of three signs that may or may not be followed immediately by a terminal.  Of the terminals, POT is the most frequent.  But I do not see these three signs forming a group with POT.
Tablet M-519A showing Yadav et al's three-sign "ender" MAN HOLDING POST / DOUBLE GRIDS
(preceded by POTTED THREE / CUPPED THREE (?) / MAN HOLDING QUOTE / STRIPED MALLET
CAGED BY BI-QUOTES / and possibly an additional SINGLE POST between).


Much the same problems arise with the three-sign combinations isolated by Yadav et al (2008: 59).  All four of the “beginner three-sign combinations” include two signs from the prefix plus the first symbol from the medial segment (or, more briefly, PM).  Only one of nine proposed “ender three-sign combinations” is not a medial plus terminal (MT) sequence, namely, MAN HOLDING POST + GRID + GRID (which Wells would classify as two signs, MAN HOLDING POST + DOUBLE GRIDS).  In 22 inscriptions in the KP concordance, MAN HOLDING POST appears without the following DOUBLE GRIDS (KP 1982: 32-33).  It is probably significant that the sequence MAN HOLDING POST + DOUBLE GRIDS does occur in 14 inscriptions.  
Interestingly, other anthropomorphs occasionally take the place of the MAN HOLDING POST.  In M-383, there is the sequence MAN HOLDING FOUR QUOTES + DOUBLE GRIDS.  Plus, in two inscriptions MAN HOLDING FOOTED STOOL precedes the two “grids” (M-161 and KP 2434).  In a good many cases, too, DOUBLE GRIDS is not accompanied by any type of “man” (I count 27 examples in Koskenniemi and Parpola 1982: 136-138).  In fact, without the aid of the computer one merely needs to glance down the columns of inscriptions in the concordance to find possible groupings: STRIPED LOOP UNDER CHEVRON + DOUBLE GRIDS (4); DOUBLE GRIDS + DEE-SLASH (4) (not to mention 13 inscriptions with STRIPED TRIANGLE + GRID, not doubled).  So, MAN HOLDING POST + DOUBLE GRIDS is a good example of a group.  But most of the other proposed beginners and enders are not.

When Yadav et al proceed to two-sign combinations, they are on firmer ground (2008: 60).  Here, they isolate five prefixes (BOAT + PINCH; FAT EX + PINCH; CARTWHEEL + PINCH; CARTWHEEL + BI-QUOTES; VEE IN DIAMOND + BI-QUOTES).  A final pair is not a prefix: VEE IN DIAMOND + STRIPED FAT LEG LAMBDA (20 occurrences cited).  Their “ender two-sign combinations” are somewhat more problematic (2008: 61).  Here, they find three terminal pairs (POT + COMB; POT + MAN; TRI-FORK TOPPED POT + POT).  But the rest combine a medial sign with a terminal – except for GRID + GRID, which could be considered a single element (as DOUBLE GRIDS).

In each grouping – four-sign combinations, three-sign combinations, and two-sign combinations – their middle combinations are more informative than their proposed beginning and ending sequences.  But since they mix prefix and terminal elements in with medial signs, the utility of the groups they isolate is limited.  They do go on to point out text beginners (which Korvink terms variable elements in the prefix), text enders (Korvink’s terminals), and auxiliary text enders (the constant in the prefix, in Korvink’s terms).  However, although they seem to recognize the same three parts of a text as Korvink, they follow what seems to be a backward sequence of steps in segmenting inscriptions (2008: 63): 

STEP 1: Search for two-sign, three-sign and four-sign texts successively.

STEP 2:  Search for frequent four, three and two-sign combinations successively.

STEP 3:  Search for Text Enders, Text Beginners and Auxiliary Text Enders successively.

That is, they first look for short inscriptions (their "texts") that repeat.  In the second step, they look for combinations of two, three, and four signs that appear within longer inscriptions.  In the final step, they look for something akin to Korvink's three segments.  Let us examine the “Text Segments” that result from this sequence of procedures.  The following derive from Table 18 with my addition of an analysis following Korvink (2008: 70):

Text No.
Combination after segmentation
Freq.
Korvink’s Analysis
1
VEE IN DIAMOND + BI-QUOTES
99
P
2
WINGED MAN + POT
57
MT
3
CARTWHEEL + BI-QUOTES
46
P
4
HAIRY HUNCHBACK + POT
37
MT
5
CUPPED SPOON + 3 POSTS + SPEAR
34
MT
6
TRI-FORK TOPPED POT + POT
28
T
7
TWO POSTS + FISH
26
M
8
CUPPED SPOON + 3 POSTS
25
M
9
BOAT + PINCH + POTTED ONE
24
PM
10
TRI-FORK + POT
24
MT
11
FAT EX + PINCH
21
P
12
QUAD-FORK + POT
21
MT
13
FAT CEE + POT
20
MT



Yadav et al have isolated three prefix pairs here (P: 1, 3, and 11) and one terminal pair (T: 6).  In seven of the thirteen combinations, though, they have crossed a segment boundary (PM in 9, MT in 2, 4, 5, 10, 12, and 13).  In only two combinations do they clearly have something (those marked M for medial segment).  These include TWO POSTS + FISH (number 7 in the list) and CUPPED SPOON + 3 POSTS (8 in the list).  Other potential pairs appear amid the middle two-sign combinations (2008: 60).  These include MALLET + FORK (shown in two forms, once with TRI-FORK, once with a symbol that combines occurrences of QUAD-FORK and QUINT-FORK); RAKE + FISH; RAKE + WINGED MAN; OVERLAPPING CIRCLES + 2 POSTS; PRAWN + ZEE; and possibly FISH UNDER CHEVRON + WHISKERED FISH.  Proposed three-sign combinations from the middle include MARKED FISH + CUPPED SPOON + 3 POSTS; PRAWN + ZEE + CROSSROADS EX; CRAB + RAKE + FISH (2008: 59).  But the significance of these pairs and triads is muted by their appearance of the many cross-segment groupings that represent spurious combinations.
Detail from seal H-24 showing Yadav et al's three-sign middle combination
MARKED FISH / CUPPED POST / 3 POSTS (plus HAIRY HUNCHBACK / DOTTED WINDOW / POT).


Why did Yadav et al come to such different conclusions from Korvink?  I think the answer lies in the first statement from their conclusion (2008: 71): “Many frequent sign combinations make their appearance as independent texts and hence considering these frequent sign combinations as units of information is justified for segmenting these texts.”  Korvink began without making the assumption that a sequence found as an independent inscription is a unit of information.  He only suggested units of information – and boundaries between them – after his analysis, as a result of that analysis.  Yadav et al, in contrast, assume before starting analysis that a short inscription is a unit of information. 
That does not sound implausible, on the face of it.  But in Korvink’s analysis, an inscription containing only two signs may consist of two segments: HAIRY HUNCHBACK + POT, for example, contains a medial segment made up of one sign (HAIRY HUNCHBACK) and a terminal (POT).  The combination might then form a single unit of information, as Yadav et al assume.  But it also might contain two bits of information, one conveyed by the medial segment, a second by the terminal.  Since meaning is not addressed in either study, this cannot be determined yet.  This indicates the wisdom of not making too many assumptions too early.

The second study to be briefly reviewed here is by part of the same team (Yadav and Vahia 2011: 1-36).  In this article they examine the signs themselves, in particular the ligatures.  Here they divide the Indus signs into basic signs, provisional basic signs, and modifiers.  The basic signs are the ones they do not consider ligatures, ones that should not be broken down further.  Modifiers are the small additions made to multiple other signs, including the four dots I term "caging" (following Wells), and the chevron placed over another symbol.  Strictly speaking, modifiers do not occur as independent symbols (2011: 9).  But of course, the CHEVRON is an independent, though infrequent, sign.  It is also possible that "caging" is an addition of STACKED FOUR to another sign. 
Tablet M-1425A with inscription that may be read various ways (from right, but left of the "tree"):
SEATED MAN HOLDING CUP / 4 QUOTES / CUP / MAN BETWEEN POSTS / COMB
(endless knot).  The tree and knot are not read, but should the "seated man" be?  Or is it pictorial?

Nevertheless, the division into types of elements is useful.  The provisional basic signs are somewhat more difficult and may actually fall into the group of basic signs or that of the modifiers.  Elements that do appear in ligatures but do not occur independently are provisional basic signs.  They include: "wobbly legs" (appears as part of one MAN HOLDING CUP, where the legs are shown bent, with small feet); "cross-legged" (the SEATED MAN HOLDING CUP, which occurs twice); "A" (the AY that usually stands over a few "quotes" or a small "comb"); "inverted box" (part of BATTERY, also part of various EXIT signs); "two-legged Y" (replacing the VEE in two examples of VEE IN DIAMOND -- essentially the two strokes fail to meet); "VA joined" (the very brief zigzag at the bottom of EN UNDER TABLE); "inverted U" (which I term ROOF); "H" (which occurs beneath ROOF); "shaded fish-like shape without fins" (the STRIPED LOOP always found beneath CHEVRON); "vertical line with a bulb at the bottom" (which I term SPOON, appearing inside various CUP or POT symbols).
I am inclined to disregard the "wobbly legs" as a rare variant form of the MAN element.  Other rare variants give one or another anthropomorph a round head, feet, or occasionally a thicker body with some hint of shoulders.  If none of these other variations is to be classified as a modifier or basic sign, I think the "wobbly legs" should not either.  The seated anthropomorph as a whole could be distinct from the standing one (although in the KP concordance, the two fall together).  But it seems odd to segment the crossed legs when the top half of the symbol also differs.  All the other representations of holding have the MAN standing with one arm on either side.  The object held is attached to one (or occasionally both) arm by a short diagonal stroke.  But the seated anthropomorph is in side view, with both arms on one side holding the object.  Thus, "cross-legged" or SEATED MAN is a separate provisional basic sign, in my view.
Bar seal H-146 with inscription: SINGLE POST / AY ON QUOTES / CIRCLED FORK / POT HATTED BEARER
(this is a common form of AY, but is it a ligature with 3 QUOTES?  Note its pairing with CIRCLED FORK, found
in 17 out of 22 inscriptions containing AY).

The AY actually does occur independently and should be classed as a basic sign, even though it is rare.  But the "inverted box" is not found independently, only as part of a more complex symbol.  Since the other parts of BATTERY and the various EXITS also do not appear independently -- for the most part -- they, too, should not be deconstructed in my view.  I think the "two-legged Y" is most likely to be the equivalent of a "hand-writing" variation.  There are other symbols found especially on the tablets or in graffiti that include strokes that do not meet or otherwise vary slightly from the more careful examples on most seals.  I view the "VA joined" or EN (3 strokes) as a variation on ZIGZAG which has two other variations, one resembling "W" (4 strokes) and one with 5 strokes, where the shortest version appears to save space in the ligature. 
However, the ROOF is probably a good element to take as a provisional basic symbol.  The independent occurrences may differ from those in ligatures, but that is a discussion for another time.  The small element beneath the ROOF that Yadav and Vahia call "H" might be BI-QUOTES; it might indeed be "H" which has an independent existence (see my post on AITCH); and it might even be a repetition of the EN.  The single example where it occurs is just not clear.  The LOOP appears in ligatures without striping, attached to BATTERY and to PRAWN; it also occurs independently in graffiti.  It makes a good provisional basic sign, perhaps with striping ("shading") added as a modifier.  But it is also possible that the more frequent appearance of STRIPED LOOP UNDER CHEVRON, a sign in which the "chevron" often lies directly on the "loop" rather than standing a bit above as in other signs, indicates this is an undeconstructable sign. 
I remain uncertain of the best way to classify the SPOON.  It is clear that it has variations as well: sometimes a "bulb" at the bottom, but sometimes a small triangle, and occasionally this little triangle is striped.  The CUPPED SPOON varies freely with CUPPED POST also, but other occurrences of the POST (alone or in ligatures) do not show any "bulb" at the bottom.  In conclusion, the proposed basic signs are basic, but the provisional basic signs need more analysis and discussion.
The modifiers are as follows (2011: 11-13):
"Vertical stroke on both sides" (when a sign is BETWEEN POSTS)
"Vertical stroke at the bottom" (when a sign is OVER QUOTE)
"Right downward tilted stroke" (often noted in my list as a TICK)
"Right upward tilted stroke" (usually a TICK or SLASH)
"Vertical stroke at the top" (the apparent numerals in the CUP and POT)
"Four vertical strokes enclosing the sign" (CAGED sign)
"Pointed hat" (CHEVRON over a sign)
"Small central vertical stroke in the centre" (DOT inside a sign)
"Slanted line on both sides" (a sign occurs between SLASHES, sometimes termed POSTS in my list)
"Slanted line in the centre" (the mark inside the MARKED FISH)
"Flat hat" (the TABLE over another sign)
"Shading" (when a sign is STRIPED)
"Four pairs of vertical strokes enclosing the sign" (DOUBLY CAGED sign)
"Angled hat" (a variation on the TABLE over a sign)
"Long angular line" (described as a SLASH inside a sign)
"Three pairs of vertical strokes and one single stroke enclosing the sign" (in WATERY SEVEN)
"Horn like attachment" (as in WINGED MAN)
"Leaf like attachment" (the sign has an EAR)
"Line with attachment to the sign" (sign with attached POST)
"Several strokes at the top of the sign" (seen in DOUBLY WHISKERED FISH)
"Several strokes at the bottom of the sign" (where CUP or POT is ON PRONGS)
Here, my terminology differs from that of Yadav and Vahia, but we see generally the same modifications.  I would not include the WATERY SEVEN, which resembles the DOUBLY CAGED marks, minus one.  That is an odd sign that needs specific attention.  I would also group some of these together ("flat hat" and "angled hat").  But this sort of minor disagreement is typical when viewing Indus signs.  Just as no two researchers find exactly the same number of signs, no two see exactly the same number of ligatures.
But Yadav and Vahia have done a service to the field by describing these divisions so minutely.  They go further and provide data on frequency of each at the major and minor sites.  Thus, one can see, for example, that "caging" is relatively frequent as a modification, found at Mohenjo daro (74 examples), Harappa (25), Lothal (12), Kalibangan (2), Chanhudaro (5), and other smaller sites (2), but not in Western Asia (2011: 15).  Figures are based on Mahadevan's sign list.
It is interesting, though, that the small FORK attached to a number of signs is not included in this list.  I imagine that is because TRI-FORK (and other variations with more strokes) exists independently.  I think if I were analyzing modifiers, though, I would include the small attached FORK despite that, just as the CHEVRON is a modifier and the SINGLE POST a modifier, even though these also have an independent existence.  It is possible that the attached forms serve a function different from that of the independent signs, especially in view of the analysis of contexts of occurrence by Yadav and Vahia (2011: 20-24).  In general, although one could view the ligature as SIGN 1 + SIGN 2, the combined sign does not appear in the same contexts as the sequence of those two independent signs.  For example, MAN BY CHEVRON seems to combine MAN and CHEVRON, a sequence that does occur in Mahadevan's concordance, as Yadav and Vahia note.  But they find "no signs common on either side" (2011: 20).  This is a valuable observation, one which advises against a facile "reading" of ligatures as a sequence of signs.


REFERENCES

Korvink, M.P. 2008. The Indus Script: A Positional Statistical Approach. Gilund Press.

Koskenniemi, K. and A. Parpola. 1982. A Concordance to the Texts in the Indus Script. Helsinki: Department of Asian and African Studies, University of Helsinki.

Possehl, G.L. 1996. Indus Age: The Writing System. Philadelphia: University of Pennsylvania.

Yadav, N., M.N. Vahia, I. Mahadevan, H. Joglekar. 2008. “Segmentation of Indus Texts,” in International Journal of Dravidian Linguistics 37 (1): 53-72.

Yadav, N. and M.N. Vahia. 2011. “Indus Script: A Study of its Sign Design” in SCRIPTA, Volume 3 (June 2011): 1-36.