Tuesday, August 31, 2010

Cee and Backward Cee

The third Indus symbol I'll be writing about looks like a parenthesis.  Most of the time, like our letter "C," its curve bulges to the left.  If it were a parenthesis, it would be the first one and you'd expect another later on, turned the other way.  But it doesn't necessarily work that way in the Indus script.  When it curves like our letter "C," I call it CEE.  When it curves the other way, like a closing parenthesis, I call it BACK CEE, which is a lot shorter than anything else I could come up with.  Either way, it has the numerical designation I3.  The Roman numeral indicates that it takes one stroke to draw it and the Arabic numeral indicates that it is the third sign in my list.  I further distinguish each variant with a letter, "a" for CEE and "b" for BACK CEE.

This symbol has two designations in Koskenniemi and Parpola's list, 154(a) and 163 for the two directions it turns.  Wells assigns a single number, 572, noting that it occurs 35 times and has only a single variant, which is odd.  There are 22 occurrences at Mohenjo daro, 10 at Harappa, 1 at Lothal, and 2 at Kalibangan.  In looking through these, I seem to see a bit of variation.  Four are BACK CEE, while the rest are CEE, for one thing.  Less significant is the degree of curvature.  Some are nearly straight, while others are more like the parenthesis in their curve.  A couple are more angular, while one is definitely pointed and thus more of a "greater than" sign.  In addition, one has a bit of a round bulge in the middle, which changes it into another sign, the BOAT (to be discussed in another post).

Fairservis calls this F-7 when it is a CEE, considering it a representation of the crescent moon.  In this case, he says, it means "great; night."  But when it is turned the other way and is BACK CEE, then it is H-3, and represents a bow without a string, or a shield.  In that case, it means "sell; price; trade."  Either way, the CEE/BACK CEE has an odd predilection for standing next to an apparent numeral, especially five.  But that's a story for another post as well.

In Egyptian hieroglyphs, there was a symbol that actually was a crescent moon and it meant a crescent moon (N11).  However, it did not stand up on its pointy end as we might expect.  Instead, it was a lazy crescent and lay on its points, rather like a low, upside-down "U."  I guess the Egyptians saw the top of their crescent moons, not the side.  Another slightly "U"-like glyph was two hills with a U-shaped valley between them (N26).  This is dw, "mountain," representing a sand-covered mountain over the edge of the green cultivated land, the way it is in Egypt.  The closest thing to a "C" shape is a fishing net (T24).  But this glyph has two loops to the side and several little bits coming off the edge, plus two smaller loops at the ends of the big loops.  So, it is not a particularly good parallel for the Indus sign.  There is the simple upside-down "U" (minus serifs) which is a hobble for cattle (V20).  But it would hardly work turned on its side to match the Indus sign.  Or there is the wide, flat-topped half-circle (V30).  This is a basket.  But then a basket isn't much better than a hobble turned on its side, is it?  I'm afraid Egyptian is not much help here.

Old Chinese has a vaguely similar sign, yi1, which means "germination."  But the similarity is not close, even in the oldest versions (Wieger 1965: 34).  The Chinese character once looked something like a sloopy "L" shape, which only a few of the straightest Indus variants approach.

In the rock art of Native Americans, there are similar shapes on occasion.  There is a CEE shape in one panel in Texas above an incomplete horse (Newcomb 1996: 118 Pl. 78).  Further west, it is easier to find similar shapes oriented horizontally, so that they are low and wide "U" shapes or upside down "U" shapes (Heizer and Baumhoff 1984: 137, fig. 74 b).

In southern Iraq, the symbols known as proto-cuneiform contain an almost identical CEE sign.  It is called SU2, as it is the ancestor of what came to be that word in Sumerian.  In that later language it meant "wisdom, knowledge."  But it may not have meant quite the same thing originally.

In ancient Iran, proto-Elamite did not have quite the same symbol.  Somewhat reminiscent of this is the rounded wedge, made by pressing the stylus into the clay so that the flat edge is upward and the rounded edge downward.  This makes a somewhat semi-circular or "C"-like shape, though oriented like the Egyptian basket.  It has the designation M344.

In Australia, two BACK CEE shapes are noted amid circles and ovals on the wall of Paroong Cave, Mount Gambier region, South Australia (Flood 1997: 91).  No meaning is known for these symbols.

The CEE and BACK CEE shapes are commonly found in the Indus script surrounding other symbols.  In some cases, the CEE precedes the other sign and the BACK CEE follows, so that the full combination very much resembles our use of parentheses.  There are at least five of this type (KP54, KP55, KP 67, KP136a, and KP136b).  There is one more example where this is done (KP53) but, in addition, a little triangular "ear" is stuck to the upper part of the BACK CEE.  This little ear shows up in other signs as well, but never on any other CEE.

In another type of ligature, the BACK CEE both precedes and follows the other sign (KP10b and KP143).  The third type is to precede the other sign with the BACK CEE and follow with CEE (KP247 and KP115).  Thus, among the Indus symbols, we see the following combinations with CEE and BACK CEE (where "X" stands for another symbol of some kind:

(X)
)X)
)X(

What could these kinds of enclosures mean?  Well, in Chinese, there's a simple, two-stroke character representing the legs of a person, ren2.  When two of these are combined in a single character so that both appear to be turned to the left, the word is cong2, "a man walking after another; to follow, to obey" (Wieger 1965: 79).  But if the one on the left is turned left (the normal direction), and the one on the right is turned right (opposite the first one), this is bei3, "note to follow each other, to turn one's back, disagreement" (Wieger 1965: 80).  The modern character no longer looks like two of the original, but that was its origin.

The point is, simply turning a curved line one way may have suggested compliance or agreement.  Turning it the other way may have suggested the opposite, disagreement or even defiance.  This is one possibility in a universe of many others!

Friday, August 20, 2010

Second Symbol: Single Post

The second symbol in the Indus script, at least in my list, is a vertical line.  I designate it I2, the Roman numeral "I" indicating that it is drawn with a single stroke, the Arabic numeral "2" indicating that it is the second in the one-stroke list.  The name I've given it is SINGLE POST, to distinguish it from the short stroke (SINGLE QUOTE).  It has other designations in the lists of other authors (KP121c, W198, Fs O-1 and K-3).  It's rather an odd thing, but according to Wells it appears 42 times in inscriptions, while it occurs 149 times according to Korvink (Wells 1998; Korvink 2007).

I mentioned the single vertical stroke in Egyptian hieroglyphs in my previous remarks (glyphs Z1).  There is somewhat similar glyph which is usually horizontal but occasionally vertical (Aa 11 and Aa 12).  It looks like two parallel lines, close together, with a short line joining them at one end.  Usually this joining line is at a slant, so that the top line is slightly shorter than the bottom line.  This is the Middle Kingdom form (Aa 11).  But an earlier form had the two parallel lines more nearly equal (Aa 12) in the Old Kingdom.  This odd glyph, which might possibly represent a platform or pedestal, was a phonetic indicator.  It stood for three consonant sounds transliterated m3', in which the last two are types of glottal stops something like the sound in the middle of English "uh-oh."

King Tut's names, showing three of the vertical strokes
In Old Chinese, a single horizontal stroke was and still is the numeral "one," now pronounced yi1.  There is also a vertical stroke, gun3, "perpendicular" (Wieger 1965: 31).  In the character for "tree," this vertical stroke represents the trunk; in the character for "center, middle," this same vertical stroke represents an arrow; and in the character shen1 "to gird up," according to Wieger, this very same vertical stroke represents a man.  What a useful mark!

In Luwian hieroglyphs (used in Turkey during the Bronze Age, and later in Syria as well), a slanting stroke was used as a phonetic symbol.  What we would call a backslash represented the syllable ra or ri.

In the rock art of North America, a single vertical stroke can be seen at times.  One appears between a small circle and a short upside-down "U" shape, over two dots (Newcomb 1996: 196, Pl. 147 no. 24-G).  Another appears between a large deer with antlers and a small deer with an object resembling a pot on its back (Pl. 147 no. 24-H).  Yet another appears to the left of a long-eared (or long-horned) quadruped (Heizer and Baumhoff 1984: 126, fig. 63 a).  A horizontal stroke can also be observed.

In proto-cuneiform, a single horizontal stroke is designated N57, the initial letter of this designation indicating that it is a numeral.  There is a virtually identical mark in Proto-Elamite, M001, which may or may not also be a numeral.  A variant, M001~b, is the same line with a wedge impressed at one end.  This may be a distinct symbol.

In Linear B writing, a single vertical stroke is sometimes used as a nonlinguistic mark to separate words.  Similarly, as noted in the previous post, the Egyptians used their vertical stroke most often to indicate that a given glyph was functioning as an ideograph rather than as a determinative or phonetic symbol.  Thus, the relative frequency of the single stroke in Indus inscriptions may not indicate meaning in the sense of a grammatical marker.  The SINGLE POST may have served a function similar to that of the Egyptian or Linear B stroke; i.e., it may have served as a nonlinguistic indicator of some kind. 

The same is also possible for the SINGLE QUOTE.  This sign is more often cited as an example of a suffix due to its frequency.  However, positional regularity is not, in and of itself, proof of linguistic function. 

Egyptian determinatives are an example of this sort of thing.  These symbols were not pronounced at any time.  They provided clues for the reader about the type of word.  This was necessary for comprehension because the nature of the writing system did not provide full details on the pronunciation of words.  Consonants were indicated but not vowels, so there was a certain amount of ambiguity built into the system.  The determinatives helped clarify things.  Now, the determinatives' position was quite regular, always coming at the end of the word.  If we examined the position of the seated man -- a common determinative -- we might be tempted to consider this glyph a grammatical suffix due to its final position in so many words.  But it is not a suffix.  It has no grammatical function when it is a determinative despite its position at the end of the word.  It has a function in the writing system that is essentially independent of language.  So it may be in the Indus script, particularly if that script is not closely coupled to language as I consider likely.

In the picture above, the reader can see another use of the vertical stroke in Egyptian.  Three vertical strokes indicated the plural, which was pronounced w.  Eventually, the triple strokes could be used as a phonetic symbol, indicating this sound.  Earlier, only the chick represented this sound.  In King Tut's names, both symbols are used.  The name most of us are familiar with, Tutankhamen, includes the chick.  His less familiar name, Nebkheperure, makes use of the three vertical strokes.  We can observe another oddity of Egyptian in both names, the honorific placing of the divine portion of the name in initial position.  In "Tutankhamen," the divinity so honored is Amun, so the reed that resembles a feather starts it off (i), followed by the side view of a gameboard with all the little men on it (mn), and in case you didn't get it from that there is a zigzag representing water (n).  It's the middle section of this name that spells out the "Tut" part, the two little half circles which are bread loaves being the "T" consonants and the chick being the "w" in between.  The "T" shape with a loop on top is the ankh, the symbol meaning life which originally came off of a sandal.  That's the middle of Tut's name. 

That leaves three more symbols at the bottom of Tut's cartouche, doesn't it?  They are not part of his name, but they are inside his cartouche quite regularly, another oddity of the ancient Egyptians.  They tell the ancient reader in the know that he was king of Upper and Lower Egypt (in case you had any doubt).  The other cartouche begins with the other divinity's name, Ra.  That's the big circle, which should have a dot in the middle.  But it's hardly going to be confused for any of the other circular signs in Egyptian when it's in front like that, so it doesn't really matter that they've out that little detail.  The part that was pronounced first, Neb, actually comes last in this cartouche.  That's the large half circle representing a basket.  It wasn't a deity and impressed no one.  The second element in the cartouche was also a deity, the dung beetle Heper (or Kheper to be a bit more accurate about the pronunciation).  It was the beastie who rolled the sun (Ra) across the sky, so it had to come next to the sun's disk in the writing.  And somehow linguistically the three little posts had to come next because there were really many little divine dung beetles, one for every day evidently, and that's what the three posts represented.  So they got to go next, pushing Neb the decidedly mortal basket to the tail end.

All of which goes to show that when one is dealing with an unknown script, all bets are off.  The Indus folk might have been following all sorts of arcane ordering schemes that had to do with honorific rules, nongrammatical clues to meaning (similar to determinatives), nonlinguistic separator signs, and so on.  Simply declaring that a sign is an ending because it comes at the end is too easy!  Besides, as we'll see later as we go along, some of the symbols that are called endings show up all alone.  One never sees that in real endings.  You never see someone get a ring inscribed "-ed."  Nobody ever got a tatto that said "-ing."  Nobody I know anyway.

Indus Script Sign One: Single Quote

I'm starting this blog as part of a quixotic quest to put a little order into one end of an academic field very few people are interested in.  There aren't many people who get excited about linguistics, you see.  Most don't even know what that is.  In this morning's paper, linguistics was blamed for a little mix-up about some bridges between Texas and Mexico, in fact.  The real culprit is semantics, which is only one branch of linguistics.  The problem, it seems, is that the officials in charge of these bridges don't call them bridges because the structures were built to control the Rio Grande, not to enable people to cross it.  But cross the river people do, as demonstrated by footprints on the structures.  And that's what's gotten people's knickers in a twist and made them give linguistics a bad name.  Bad, bad linguistics!  If only those officials had called those structures bridges right from the get-go, those river-crossers wouldn't have crossed that river -- or so the argument seems to imply.

Well, that's neither here nor there.  I'm just a mild-mannered linguist and I'm perfectly willing to call a bridge a bridge and let anybody cross it who wants to.  My real interest lies elsewhere -- in the Indus script.  Back in the Bronze Age, on the shores of a very different river, some folks were using symbols and not leaving notes about it that we can read today.  Their river is now known as the Indus, and archeological remains of their civilization reside in museums in India and Pakistan.  But no one can read their script and no one knows for sure what their symbols mean.  Plenty of people have made attempts and any number of claims of success have appeared in print and on the internet over the years.  It's even conceivable that somebody is correct.  The only problem with granting success to one of those decipherers is this.  Since all those successful decipherers disagree with one another, how does the observer know which one is right?  Who really deciphered it if each one says he's right and all the others are wrong?  Personally, I'm not impressed by the loudest screamer or the most popular charmer.

Let's back up a bit and try something a little more scientific.  How many signs are there in this script, to start with?  Unfortunately, there isn't even an agreed-upon answer to this question.  Depending upon which expert one asks, one discovers that there are anywhere from about 200 signs to about 700 signs.  That's rather a large difference.  Various scholars have published lists of these symbols from time to time, including Asko Parpola, Iravatham Mahadevan, Bryan Wells, and Walter Fairservis.  Some have published concordances as well, listing all of the inscriptions.  In addition, there are two volumes of photographs showing the actual seals, tablets, and other artifacts, in case you want to puzzle out the information for yourself.  A third volume is apparently about to be published or has just recently come out.

A group of scholars has recently attacked the very idea that this is a writing system (Farmer, Witzel, and Sproat 2005).  Their argument is complex and covers a number of angles.  No single point is altogether convincing by itself, but putting them all together makes a pretty good case.  Some of the quicker points include these observations:  there are no long texts in this script, only short ones; there are no depictions of scribes on archeological artifacts; the script did not develop a cursive style; there are no monuments with inscriptions; and as more seals and things are discovered, more signs are discovered.  So the number of signs is increasing rather than decreasing, as time goes on.

Still, even if we assume that this was not true writing, the symbols must have had some sort of meaning.  And fully developed writing systems -- like Egyptian hieroglyphs and Akkadian cuneiform -- are not the only types of symbol systems known.  In his review of the history of writing, I.J. Gelb discusses several types of symbol systems that aren't quite writing (1963).  He groups them all under the term "semasiography."  This coined expression apparently didn't catch on, which is unfortunate.  Instead, one usually sees the clumsy expression "proto-writing" for the systems that preceded cuneiform and hieroglyphs.  That's good enough for a system that eventually evolved into a fully developed writing system.  But next door to "proto-cuneiform" in ancient Iran, there was a symbol system that was quite similar.  It was meaningful, it was complex, but it was not closely tied to speech.  But it did not evolve into writing.  Instead, it died out.  It is known as proto-Elamite.  But it is most likely not an ancestor of the historical Elamite writing, despite that name.

Proto-cuneiform is comprised mostly of symbols that appear only once or a few times, far less than 100 times apiece.  These are thought to represent owners, in some sense.  They may be individuals or institutions, such as temples or palaces.  Aside from this large set, there is a small set of signs that appears very often.  These are better understood.  They represent mostly commodities, with a few indicating some quality or qualification of the commodity.  There are also numerical symbols.  But, to begin with, there is not a clearcut distinction between numeral and commodity.  To represent a concept such as "6 sheep," a scribe might well duplicate the sheep symbol six times rather than write one symbol for "six" and one for "sheep."

The situation appears to be much the same for Proto-Elamite, although this script is understood much less well than proto-cuneiform.  The vast majority of signs appear only once or a mere handful of times.  Only a relatively small percentage of symbols occur repeatedly.  These appear to mostly represent commodities.  Interestingly, they are often the same commodities found in proto-cuneiform: livestock, bread, beer, grain, and humans of low social status.

Now, nothing is entirely clear as yet about the meaning of the symbols in the Indus script.  But statistics on the appearance of symbols have been published and they are quite similar to those for proto-cuneiform and proto-Elamite.  The majority of symbols appear once or only a few times.  A small core of symbols appear more often, often enough that statistical study can be made of them.  This does not prove that the core symbols are commodities in this case.  But it does suggest that the Indus script is another proto-writing system, not a fully developed writing system.

That suggests that there probably never were any long texts on perishable materials, despite what some researchers have supposed.  The Indus people were capable of writing longer strings of symbols on pottery, or on clay tablets, just as their western neighbors did.  But there is no evidence that the people of the Indus Valley ever did this.  There are only seals, metal tablets, bangles (and a few other items), and pot shards with fairly short inscriptions.  These inscriptions have from one symbol to 23 or so -- the exact count varies from one source to the next.  But no inscription has as many as 30 symbols.  That much is clear. 

The fact that the Indus script only appears in short inscriptions means something.  The people chose to do this.  That doesn't make their symbols or their inscriptions defective.  That's why I think "proto-writing" is an unfortunate term.  That's a useful term when you're studying the history of writing.  But not everybody intends to develop writing.  Sometimes people use symbols for other purposes.  For example, if we look at modern Navaho sand paintings, we see a great variety of symbols which have clear purposes, clear meanings, and whose placement is anything but random.  These are not writing and they are not proto-writing.  They aren't meant to be anything like writing.  These kinds of symbols serve an altogether different sort of purpose.

I once had a friend from Brazil who had an emblem that she hung on the inside of her front door.  It was round and bore various markings that I recognized as letters from the Hebrew alphabet and characters in Chinese.  I asked her what the Hebrew letters stood for.  She responded that she was unaware that they were Hebrew letters.  I then asked why the Chinese character for good was drawn in such a curious way.  This character is made by combining the characters for "woman" and "baby."  But in this case the "baby" portion was missing its arms, which is the way one normally writes a rather different character, one which would not combine with "woman" to make the word "good."  She was equally unaware that this was a Chinese character and also unaware that the "baby" was defective.  The whole thing was a good luck charm and each component was also a good luck charm.  Its purpose was to keep her house safe and its placement on her door was to keep bad influences from entering her abode.  The moral of the story is, what looks like writing may not actually be writing.  It depends on why something is made, who made it, and who uses it.

I am going to go through the list of Indus symbols, as I see them, giving each a numerical designation based on the number of strokes it takes to write it.  This organizational scheme derives from Chinese, a modern language which arranges characters according to radical (a root or basic form) and stroke number in the dictionary.  I am also giving each Indus symbol a name, in order to discuss them more easily.  Scholars dealing with proto-Elamite or proto-cuneiform sometimes name one or another symbol, in order to discuss it more easily.  They almost never name all of the signs in any system.  However, it seems a logical move to name all of the Indus signs to enable us to transcribe the inscriptions via a standard computer keyboard.  We could use just the numerical codes, once these assigned.  But nobody can type purely numerical codes without error and these are very difficult to proofread and correct.  Names are, by their very nature, more susceptible to such correction.  Hence the duplication.

The first Indus symbol in my list resembles a single quotation mark.  I give it the numerical designation I1.  The Roman numeral indicates that it takes a single stroke to write it.  The Arabic numeral indicates that it is the first in the one-stroke list.  Its name is SINGLE QUOTE, written in all caps to indicate that it is, in fact, only a name and not the meaning of the symbol.  It so happens that there are two variants of this symbol, which can be distinguished by position.  The high variant we can indicate with a tilde and the letter "a," as is done to designate proto-cuneiform and proto-Elamite variants (SINGLE QUOTE~a).  The mid-level variant we can indicate with "b" (SINGLE QUOTE~b).

Kimmo Koskenniemi and Asko Parpola published a list of Indus symbols and a concordance (1982).  We can designate this sign's position in their list thus: KP121.  Bryan Wells compiled a more thorough list as part of his master's thesis (1998).  We can designate the same sign's position in his list thus: W196 (82 occurrences).  Walter Fairservis also gave this sign a designation, FsP-1 (1992). 

Michael Korvink analyzed the core signs of the Indus script and determined that SINGLE QUOTE functions as part of the prefix (2007).  That is, it tends to occur near the beginning of inscriptions, but not in initial position.  It causes another sign or signs to precede it in initial position.  This preceding position is a variable in the prefix, filled by a number of different signs.  The SINGLE QUOTE is a constant, coming at the end of the prefix.  This whole construction is then followed by the rest of the inscription, hence its designation as a prefix.

This sign is similar to certain symbols found in the scripts of other lands.  In Egyptian hieroglyphs, there is the glyph designated Z1 in Sir Alan Gardiner's list (1976: 534-535).  This is the single vertical stroke, used as an ideograph or determinative to indicative the numeral "one."  It also follows any glyph used as an ideograph to let the reader know that it is an ideograph and not a phonetic glyph.  It is occasionally a space filler, too, since the Egyptians were terribly fond of symmetry and neatness in hieroglyphic writing.  It could also appear in place of human figures on occasion in places where the correct glyphs were thought magically dangerous.

In Old Chinese, there was a dot, a character now pronounced zhu3 which means "point, as of flame" (Wieger 1965: 30).  This dot was first drawn as a small, round, black circle or else as a tear-shaped mark.  Now it is an oblique stroke.

In the rock art of Texas, in a scene including men, animals, and various large tally marks, there is a dot or very short stroke (Newcomb 1996: 94, Pl. 53, no. 1).  Similarly, a collection of rock art from Nevada and eastern California shows a short oblique mark below the left foot of a horned anthropoid figure (Heizer and Baumhoff 1984: 194, fig. 131 g).

In proto-Elamite, a single round mark, made by pressing the round end of the stylus into the clay, was a numerical symbol.  It was equal to six of the previous symbol (the long wedge placed horizontally), which in turn was equal to five of the previous symbol (the short wedge placed vertically).  Or perhaps it's the other way around and I have it backwards!  (see http://cdli.ucla.edu/pubs/cdlj/2005/cdlj2005_003.html ).

In proto-cuneiform, a single short wedge, placed horizontally, essentially represents the number one.  The impressed circle (the same seen in proto-Elamite) represents one of a larger unit (see http://www.cdli.ucla.edu/tools/SignLists/protocuneiform/archsigns.html ).

Thus, the SINGLE QUOTE does not seem to be a numeral in the Indus script, although it superficially resembles a single tally mark and the single mark that is the numeral "one" in other scripts.  Korvink's analysis indicates that it has some other type of function in the Indus script.  And we can see that the single stroke has such a non-numerical function in other scripts as well.  It is not always a numeral in Egyptian, although it can be.  In fact, most of the time it is not a numeral.  And in Chinese too, while a single horizontal line represents the numeral "one," a single mark of some other type is not the same thing.

Well, that's a start!

SOURCES

Fairservis, Walter. 1992. The Harappan Civilization and Its Writing: A Model for the Decipherment of the Indus Script. Leiden: E.J. Brill.

Gardiner, Sir Alan. 1976. Egyptian Grammar. (Originally 1927). Griffith Institute & Ashmolean Museum: Oxford.

Gelb, I.J. 1963. A Study of Writing. Chicago: University of Chicago.

Heizer, Robert & Martin Baumhoff. 1984. Prehistoric Rock Art of Nevada and Eastern California. (Originally 1962). Berkeley: University of California.

Korvink, Michael. 2007. The Indus Script: A Positional Statistical Approach. Gilund Press. (Amazon)

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

Newcomb, Jr., W.W. 1996. The Rock Art of Texas Indians. (Originally 1967). Austin: University of Texas.

Wells, Bryan. 1998. An Introduction to Indus Writing: A Thesis.  The University of Calgary. www.collectionscanada.gc.ca/obj/s4/f2/dsk2/ftp03/MQ31309.pdf

Wieger, Dr. L. 1965. Chinese Characters. (Originally 1915). New York: Dover.