What is this about?

Cuneiform is the oldest true writing system in the world. Developed in ancient Mesopotamia around 3,000 BCE, it was used to write everything from law cases to epic poetry to astrological omens to customer service complaints and adapted for a dozen different languages before disappearing from the historical record in the second century CE.

It’s also infamously difficult to learn.

Most instructional resources on these languages avoid touching cuneiform for as long as possible, in an attempt to not scare students away—which is a shame, because the writing system is part of why these languages are so fascinating! Thousands of clay tablets have been excavated from across the Near East, many still untranslated now. But so far, nobody has made a good introduction to the cuneiform writing system that will help people pick it up and get started.

That’s what this document is meant to remedy. It focuses specifically on Hittite cuneiformAnd even more specifically on the later (“New Script”) forms of the signs, whenever there’s a difference. These are the more common ones, since most Old Hittite tablets we have are later copies in New Script.

, used from the 17th to the 12th centuries BCE in Anatolia; Hittite uses a smaller inventory of signs than Assyrian or Babylonian, while still being fairly well understood. This document will start by covering the basics of cuneiform and how it’s written, then touch on some details of the Hittite language, before laying out a full overview of the writing system and how it encodes Hittite words. Each heading is intended to be treated as one lesson, to avoid overwhelming students with too many signs all at once.

Lesson 1A: Wedges on clay

“Cuneiform” literally means “wedge-shaped”, since every cuneiform sign is made up of wedge-shaped impressions in a clay tablet. The oldest ancestors of cuneiform involved drawing pictures into the clay with a sharp implement, but this tends to get slow and messy, with clay piling up at the end of the stylus. The solution was to press the stylus down like a stamp instead. And as a result, cuneiform is a fundamentally three-dimensional writing system.

The original styli were cut from reeds, and had roughly rectangular ends. So the distinctive wedge shape comes from the corner of this rectangle:

In the modern day, the disposable chopsticks you’ll find at many Asian restaurants work perfectly. Just make sure to get the ones with rectangular ends instead of round ones. Very early scribes did use round styli as wellYou’ll find some discussion of this in a later lesson.

, but “standard” cuneiform was always written with rectangular ones; you won’t get crisp, sharp wedges without it!

Unfortunately, this lesson is being displayed on a flat screen, which makes it difficult to show the three-dimensional wedges. Photos don’t tend to show them very well, as you can see hereSpecifically, this is a New Script form of the sign ik, the third sign on line 15 of the tablet KBo VI.1, part of the Epic of Gilgamesh.

:

So for two-dimensional screens and pages, we tend to use drawings (“autographs”) instead of photos.

You’ll find more details in lesson 1B. But for now, the only thing that’s important is to recognize what wedges look like in these drawings.

Horizontal and vertical

The most common types of strokes are the horizontal and vertical, and they’re written with two different edges of the stylus.

This means that a scribe can switch between them easily, without needing to rotate their wrist ninety degrees every time. It also means that horizontal strokes can be any length, but vertical strokes are limited by the size of the stylus. This led to cuneiform signs all having approximately the same height vertically, but varying widths horizontally, which makes them line up nicely in rows.

In Hittite, the head of a horizontal stroke is always at the left, and the head of a vertical stroke is always at the top. In other words, strokes always point down or to the right, never up or to the left. In some other styles of cuneiform (like Old Babylonian), there are sometimes upward vertical strokes, but never leftward horizontal strokes. An upward vertical just requires tilting your hand a bit more, but a leftward horizontal would require flipping the stylus all the way around!

Diagonal

The diagonal strokes are less common in Hittite, because they required actually rotating either your hand or the tablet! They’re fundamentally the same as the horizontal and vertical strokes you’ve already seen, just drawn on an angled tablet.

This means that we often find intricate arrangements of horizontals and verticals, and intricate arrangements of diagonals, but usually not arrangements that mix the two of them. It would be a pain to keep rotating the tablet back and forth, so usually all the diagonal strokes in a sign are grouped together.

The only diagonal strokes found in Hittite are the “upward” diagonal, from bottom left to top right, and the “downward” diagonal, from top left to bottom right. Just like with horizontals, diagonals never go from right to left in Hittite use—it would require flipping the stylus all the way around, and that’s just too much of a pain.

Winkelhaken

The last type of stroke is the “Winkelhaken”, German for “angular hook”. (They can also be called “hook” strokes if you want to stick to English.) These look like broad, flat triangles, and are made with the corner of the stylus.

In some styles of cuneiform, the diagonal strokes were dispensed with, and all signs were made up of horizontals, verticals, and hooks. But in Hittite, you have all five to deal with. If you ever see something that looks like a horizontal with the head on the right, it’s actually a horizontal with a hook at its tip.

Sometimes the difference between downward diagonals and hooks isn’t entirely clear, so some authors group them together as the same thing. But hooks could be written without rotating the tablet, so they’re often written in horizontal rows and vertical columns (like in the sign ce), while diagonals required tilting everything, so they’re usually stacked together diagonally (like in the sign hi).

Since this distinction is so hard to see, though, most sign search tools will offer an option for “I don’t know if this is a diagonal or a Winkelhaken”, or even have that as the default! It matters more for writing tablets than for reading them.

Exercise 1A

Get some clay (Play-Doh works well for this) and a stylus (disposable chopsticks with rectangular ends are perfect) and practice making the five different types of strokes. Then try to draw these signs:

Lesson 1B: Strokes on paper

Of course, in the modern day, most of our writing materials are flat. Students are taking notes on flat paper instead of clay, instructional materials are published in flat books, and these pages are being shown to you on a flat screen.

This is a problem, since it makes it very hard to display three-dimensional wedges! Cuneiform doesn’t photograph very well; it’s easy to see the wedges when you can rotate the tablet in your hand, but a static photo is a different story.

As a result, a lot of different conventions have arisen for how to represent cuneiform on paper and screens. Most modern scholars aren’t working with the physical object, or with a photograph; instead, they use “autograph” drawings, and the first step in analyzing a tablet is to make one of these.

Ideally, an autograph needs to show the position, length, and direction of each stroke, and where the head is. We can’t just draw the heads, because these three are different signs (the sounds qa, me, and bar)This is specifically the older form of the sign qa; the newer form, which you’ll be learning in a later lesson, has an extra stroke.

:

But we also can’t just draw the tails, because these two are different signs (the number 2 and the sound a):

So we need to somehow indicate the heads, and the tails. But there are several different ways of doing that, going all the way back to ancient Mesopotamia!

Styles of writing

Because while cuneiform was usually stamped into clay, that wasn’t always an option. Sometimes an ancient king would want to inscribe their glorious deeds in stone monuments, or paint them on the walls of buildings…or a scribe’s tablet dried out before he had time to sign his name.

I’ve written more about these different styles on my blog, if you’re interested. But all of them are trying to represent both the head and the tail of the stroke, generally using three or more pen strokes to do it. This is what makes cuneiform signs look so complicated and difficult by modern standards. But it’s important to remember that each wedge is only a single press of the stylus—even if it takes three pen lines to draw, each wedge only takes a moment to form on actual clay.

Assyrian ink

Most of the ancient examples of inked cuneiform are exceptional ones: intricate works of art, or scribes improvising with a fired tablet. But at one point, there may have been a tradition of “casual” inked cuneiform as well.

One of the last of the Assyrian emperors, Ashurbanipal, had tablets collected from across the empire to build an all-encompassing library. Like any good librarian, he wanted the name of the library stamped on every tablet, to make sure they’d get returned properly. (And like any good emperor, he made sure his own name was included in that.)

Generally this was inscribed the usual way, with a stylus. But in some cases, the tablets looted from various conquered cities were already fired. When this happened, the librarians wrote the name of the library on in ink.

This inking style is only known from three tablets, since ink is a lot less durable than fired clay. But given how cleanly it’s written (compared to that Hittite scribe awkwardly scraping out triangles), there was probably a larger tradition around this. The head of the stroke is drawn with a single curved line, and the tail with a single straight line; in my opinion, this ends up being a lot more readable than forms which draw a full triangle for the head.

This is how the cuneiform in this article is displayed by default. If you don’t like this style, you can change it with the dropdown at the top of the page. Just remember that all of these styles are compromises to represent a fundamentally three-dimensional script on a two-dimensional screen; the best way to write cuneiform is always with a stylus on clay!

Exercise 1B

Take these signs in the Assyrian ink style and draw them on clayAgain we’re looking at an Old Hittite form of en here instead of New Hittite, since it makes a better demonstration.

.

Take these signs and draw them in the Assyrian ink style.

Lesson 1C: Pronunciation

Hittite cuneiform is specifically designed to write Hittite, so it can be useful to know a bit about the sounds of Hittite and how they were pronounced. Feel free to skip this lesson completely if you’re not interested; you can always come back later if you’re wondering how c or ā was pronounced, or why there isn’t any s.

Consonants

Hittite cuneiform represents sixteen different consonants: p b t d k g q m n l r c z h y w. However, these weren’t all actually distinguished in the Hittite language! The signs ta and da, for example, are used pretty much interchangeably; choosing one over the other doesn’t seem to indicate any difference in pronunciation. Even when writing in Akkadian, a language that did distinguish t from d, the Hittite scribes mix them up constantly, suggesting they couldn’t hear a difference between them. The same goes for p and b, and k, g, and q. In other words, the Hittite language had no voicing distinctionOr ejective distinction, for that matter. Q in Akkadian is an “emphatic” (probably ejective) velar stop.

.

Instead, they distinguish single from double consonants: while there’s no distinction between t and d, there is a distinction between t and tt. This holds true for almost all of the consonantsIt seems that z was always “double” and y and w always “single”. The rest could be either.

: apa- “that” is a different word from appa “back”. The word for “you” appears as zikka, zigga, and ziqqa interchangeably, even within the same text, but never as **zikaWhy not zikga, ziqka, and so on? You’ll learn why in lesson 5. The short version is that cuneiform never distinguished voicing at the end of a syllable, so ik, ig, and iq are written exactly the same way.

.

Scholars disagree on how this distinction was actually pronouncedSome suggest it was a voicing or aspiration distinction, based on Proto-Indo-European etymology and Ugaritic transcriptions, but the consensus seems to be moving toward a length distinction.

, but the leading theory is that the doubled consonants had a longer duration than the single ones—pp and nn were pronounced for twice as long as p and n, like in “lamp post” versus “compost”, or “unnamed” versus “unaimed”.

Since there was no distinction between the pairs t/d, p/b, and k/g/q, dictionaries alphabetize them under t, p, and k, respectively. Really, the only reason why d, b, g, and q are used at all is to make transcriptions more legible: we want to indicate the exact signs used to write a word, and ba is just easier to read and write than pa₂. Similarly, y and w are alphabetized as if they were i and u, since it’s not entirely clear if there was a distinction there, or if i u just become y w before vowels.

C, Z, and H

Most of the consonants are pronounced about as you’d expectThe same as their IPA values, for the linguistically-inclined, except that y is IPA /j/.

—we don’t know precisely how the Hittites spoke, so an English “d”, a Spanish “d”, and a Mandarin “d” are equally valid ways to pronounce a Hittite dAs are an English “t”, a Spanish “t”, and a Mandarin “t”, for that matter—remember, there’s no difference between t and d in Hittite.

. But three of these consonants are less obvious.

The c sound is written as š in most books, and was probably pronounced like an English /s/, as in “sip”In other words, some sort of sibilant fricative, and Egyptian transcriptions suggest an alveolar one.

. Some languages that used cuneiform distinguished s from c, with one being the /s/ sound in “sip” and the other the /ʃ/ sound in “ship” (see below), but Hittite never did, so both pronunciations are valid and won’t create any confusion. Hittite scribes chose to use the c signs to write their sibilant, abandoning the s signs completely, so we transcribe it that way for consistency.

In these lessons, I’ve used c instead of š simply to make it easier to type, but š is what you’ll find in published dictionaries. Some books on Hittite use s instead for simplicity, but that is not a good idea if you’re doing anything with cuneiform, for reasons that will be explained in a later lesson. Still, Hittite dictionaries alphabetize it under s, and it’s the best way to transcribe an /s/ sound in foreign names. Some people write it as sh as a compromise, but this creates its own set of problems: how are you going to write the distinction between c-h and c-c (as in echar eccai “she spills blood”) without it becoming unreadably messy?

The z sound is a /t͡s/Linguistically, it’s not clear if it was one phoneme or a sequence of two, since those cuneiform signs are also used when t and c end up next to each other for any reason. But it’s convenient to assume it’s a single one, since it’s written specially, and it can appear where other consonant sequences can’t.

, as in “pizza” or “Mozart”, or the end of “cats”. Unlike in English (but like in Russian and Japanese: tsar, tsunami), this sound can appear at the beginning of a word in Hittite, as in zik “you”. The sequence tc was always written as z, so if you like the “ship” pronunciation of c, you can pronounce z like the “ch” in “church” for consistency. This is also the best way to transcribe sounds like the aforementioned /t͡ʃ/ in “church” and the /d͡ʒ/ in “jump” in foreign names.

Finally, the h soundThe most interesting sound for historical linguists, since it’s the reflex of Proto-Indo-European *h₂, and sometimes *h₃.

is some kind of fricative in the back of the mouth. Most scholars pronounce it like the sound in German “ach” or Hebrew “chet” (IPA /x~χ/), though evidence from Ugaritic suggests it might also have been pharyngeal like Arabic “ḥal” (/ħ/) or voiced like Spanish “amigo” (/ɣ/), depending on the environment.

Any of these pronunciations works, as does the /h/ sound in English “hat”. In formal publications it’s generally written ḫ, or sometimes xThough I really don’t recommend x. It’s the IPA symbol for this sound, which is why it’s sometimes used in linguistic works, sort of like writing c as “s”, since that’s how it’s pronounced. But in cuneiform transcriptions, lowercase x on its own means “there’s a sign here but it’s too damaged to read”, the times sign × means “this sign is composed of one sign inside another”, and the subscript ₓ means “we’re confident in this reading but it hasn’t been officially acknowledged by a sign list yet”. Technically, none of these create ambiguity with using x for a consonant, but it’s just asking for trouble. So “s” and “x” are fine if you’re only talking about the language, but if you’re doing anything at all with the cuneiform writing system, use c and h (or š and ḫ in more formal publications) instead.

, but using the letter h doesn’t cause any problems, since there’s no other “H-like” sound in cuneiform to confuse it with. No matter how you choose to pronounce it, make sure it’s distinct from k, and that you can pronounce it both at the start of a word (hanti “separately”) and in the middle of one (mahhan “when”)It can’t appear at the end of a word in Hittite, but it does appear at the end of some Sumerian logograms like MAH “great” too.

.

When two consonants appear next to each other in a transcription, they do not interact with each other. Ch in Hittite transcriptions is the sound of c followed by the sound of h, like in the middle of “grasshopper”. Echar “blood” is pronounced something like “ess-khar” (/es.xar/), not *“eh-char” (/e.tʃar/).

Labiovelars

Hittite also had two additional sounds which didn’t have dedicated signs in cuneiform. These are known as the “labiovelars”, and linguists write them as k^w and h^w, as in ek^wzi “she drinks” and tarh^wen “we overpowered”; they were pronounced something like Hittite k and h crossed with w. In other words, k^w is approximately the “qu” in “queen”; h^w doesn’t have a close analogue in English, but if you pronounce “wine” differently from “whine”, the “whine” sound (/ʍ/) is a fine approximationLinguistically, it’s a labialized version of h, so probably something like /xʷ~χʷ/.

.

In cuneiform, k^w is written as either ku or uk, whichever the scribe preferred at the moment, while h^w is written as either hu or uh—and as a result, that’s how they tend to be written in our modern transcriptions as well. So outside of specialized linguistic works, these words will be written ekuzi/eukzi and tarhuen/taruhen instead, without the special ^w; this is how you’ll see them in the later lessons.

Vowels

Hittite had five vowels, a, e, i, o, and u. Some scholars don’t believe there was a distinction between o and u, and as a result, o is always alphabetized as u in dictionariesAnd often written that way, too! Most dictionaries will list ōp- “to rise” as ūp-.

, but consensus is slowly shifting toward them being distinct. The pronunciation was along the lines of the a in “father”, the e in “forte”, the i in “machine”, the o in “code”, and the u in “tuba”—that is, the vowels are spoken the way they’re pronounced in most other languages (or in the International Phonetic Alphabet), not the way they’re generally pronounced in EnglishWhen we can’t even reach a consensus on whether o and u were separate sounds, it’s hard to say anything more precise about the pronunciation. Some people prefer to transcribe as /ɛ ɔ/ instead of /e o/, for example, but in my opinion there really isn’t enough data to make judgements like that, so I’m defaulting to the symbols that are easier to type—it’s like how saying “it’s around 98.6” instead of “it’s around 100” can make people think you have more precise information than you really do.

.

These vowels are sometimes written doubled in cuneiform, just like the consonants. Unlike with the consonants, it’s not clear if this actually indicated a difference in pronunciation. It might indicate a longer vowel or some kind of stress, but it could also just help make the writing system less ambiguous, as you’ll see in lesson 4. In a few cases it distinguishes one word from another, like maan “if” and man “maybe”, but this is rare. Kloekhorst proposes that it indicated a glottal stop, but this isn’t widely accepted.

While doubled consonants are written double in transcriptions, doubled vowels usually aren’t—the Hittite scribes were less consistent about using them, and they look less aesthetically appealing to English-speakers. Instead, a barCalled a “macron”, Greek for “long”.

is put on top of the vowel if it’s written double: ā, ē, ī, ō, ū (so man versus mān). More detail can be found in a later section.

Sometimes the a vowel is even written tripled, in a very small handful of words. Based on a scribal convention in Akkadian, and the etymology of these words, it might be an abbreviation for the sequence ayya…or for whatever that sequence turned into later on. Examples are given below.

Exercise 1C

Try to pronounce these sentences in Hittite. This comes from a myth about a conflict between the storm god Tarhont and the great serpent Illuyanka. After Tarhont is defeated by the serpent in a fight, he asks his daughter Inara (the goddess of wild animals) for help; she prepares a trap for the serpent, and recruits a mortal named Hopaciya to help her in her plan.

Nu hōman mekki hantāit.
She prepared lots of everything.

Wiyanac palhi, marnuwandac palhi, walhiyac palhi,
The container of wine, the container of marnuwand, the container of walhi,

nu palhac andan iyada iet.
she put a lot in all of the containers.

Nu Inarac Ziggaratti pāit.
Inara went to the city of Ziggaratta.

Nu Hōpaciyan antuhcan wemit.
She met the human Hopaciya.

Lesson 2: Transliteration and V signs

Signs and readings

Most cuneiform scholars don’t actually use the signs themselves in their publications. Until recently, typesetting actual cuneiform signs was near impossible, and even nowadays it’s more difficult than it should be.

Instead, it’s standard to give a transliteration: writing out the pronunciation of each sign rather than drawing the signs themselves. The sign i is pronounced i, for example, so we represent it simply as i. Words are separated by spaces, and signs within a word are separated by hyphens, so a-i is transliterated as a-i.

Each “reading” corresponds to one and only one sign, and sign dictionaries list all the applicable readings, so that everyone can be sure that i specifically means i. In these lessons, you can also see the appropriate reading for any sign when you hover over it. But there are three difficulties that can trip you up.

First, each reading corresponds to one sign, but one sign can have multiple readings! The sign ki, for example, can represent either ki or ke. The scholar making the transliteration needs to pick and choose the appropriate readings to make sense of the text. This becomes an especially difficult task when logograms get involved.

Second, sometimes multiple signs can be pronounced the same way. In Akkadian, for example, the signs u and u₂ are both pronounced u. It’s important to represent this difference in the transliteration, so that readers know which signs are being talked about—otherwise they’d have to go back to the original tablet to check, which defeats the whole purpose of using a transliteration! So numbers are added to the readings to distinguish them: u is transcribed as u, u₂ as u₂, u₃ as u₃, u₄ as u₄, and so on.

This can sometimes lead to a lot of numbers in a text, so there’s a shorthand: an acute accent on a vowel stands for the number 2, and a grave accent stands for the number 3. Since every cuneiform sign used in Hittite has a vowel in it somewhere, this works out nicely. But for numbers higher than 3, there’s no choice but to resort to the numbers. In other words, the signs u ú ù u₄ are usually transliterated as u ú ù u₄.

In Hittite, with only a couple exceptionsIn particular, ur and úr, and cu and cú, are used interchangeably depending on the scribe’s whims.

, there’s only one sign for each pronunciation. But, these readings have to be unambiguous across all languages that use cuneiform—if there’s a passage in Akkadian right in the middle of a Hittite text, for example, it would be confusing to change transliteration systems halfway through! It’s the same reason we have to use c instead of s for the Hittite /s/ sound. Across different eras and dialects of Akkadian, for example, two different signs were used for the pi sound; Hittite only uses one for that purpose, but through random historical accident, it’s the one that was assigned the reading pí (that is, pi₂). As a result, the reading pi never appears in Hittite, only ever pí.

Bound transcription

Transliteration is separate from bound transcription, which just attempts to capture the sounds of a Hittite word in an easy-to-read way. A transliteration of the Hittite word for “water” might be wa-a-ta-ar or wa-a-tar, depending on the scribe’s preference; a bound transcription would be wātar, ignoring the details of sign choice to focus on the actual sound of the Hittite word.

Whether something is a transliteration or a bound transcription should generally be clear from context.

Vowels

As a result, the signs for the five vowels of Hittite are these. They’re known as the “V” signs for short.

Note again that the /u/ sound is always written with the sign ú! The sign that took the reading u is used in Hittite for /o/.

The reason for this confusing state of affairs goes back to the origins of cuneiform. Originally, cuneiform signs were used for entire words: a meant “water”, which was pronounced as a in Sumerian; u meant “ten” and ú meant “crops”, both of which were pronounced as uLong ago, “ten” may have actually been yu, but by the Akkadian period they were well and truly identical.

.

Eventually, these signs were repurposed for their pronunciations instead of their meaning, like how English-speakers sometimes write the word “too” with the symbol for the number “2”. And as a result, when the Akkadians adapted the writing system from the Sumerians, they ended up with two separate signs for the u sound, and used them more-or-less interchangeably.

When the Hittites adapted the writing system from the Akkadians, this worked out great for them—because Hittite, unlike Akkadian, had an o sound separate from u! So they used u for the o sound, and ú for the u sound. But since the sign readings have to be consistent across Hittite, Akkadian, and Sumerian (as well as Hurrian, Hattic, Elamite, and others), we have to transliterate them as o and ú, for consistency. Even if o is never used for the /u/ sound in Hittite, the name u is still assigned to it from Akkadian, and isn’t available for ú.

The five vowel signs are some of the most common in all of Hittite. They’re used whenever a vowel makes its own syllable with no consonants around it, like in the word au “look!”: a-ú. But they can also be used in “plene” spellings, which you’ll discover in a later lesson. Make sure you learn them well!

Exercise 2

Practice writing the five vowel signs, ideally on clay as well as on paper.

Adjust the style settings at the top of the page until the signs below are comfortable for you to read, since you’re about to encounter a lot more of them!

a e i o ú

Memorize these five signs, in whatever way works best for you. (Personally, I use Anki flashcard decks.) You’re about to encounter a lot more of them, and you’ll want to have a good framework for it.

Lesson 3: CV signs, part 1

Cuneiform is a syllabic writing system. Unlike an alphabet, it doesn’t have a sign for each individual sound. Instead, a cuneiform sign represents a syllable, or sometimes a piece of a syllable.

In particular, a cuneiform sign can represent a vowel on its own, but not a consonant on its own—after all, a vowel on its own is a valid syllable, but a consonant on its own isn’t. So there are signs for ta, te, ti, and tu, but no sign for just t; you can’t have a t on its own without a vowel attached.

You’ve already learned the V signs, for vowels with no consonants attached. The next category is the most common one, the CV signs, for “consonant-vowel”—that is, each of these signs represents a consonant followed by a vowel. While English-speakers would spell “banana” as b-a-n-a-n-a, Hittite-speakers would write it with only three signs, ba-na-na.

When one box spans multiple rows or columns, that means the same sign is used for all of those syllables. The signs di, gi, and ki also have the readings de, ge, and keSumerian reliably distinguished the vowel i from e, but early Akkadian did not, and the Hittite scribes adapted their cuneiform from Akkadian, not Sumerian.

. The sign bu also has the reading pu.

The sign bi also has the readings bé, pí, and pé—the signs with the unaccented readings be, pi, and pe are used for totally different things in Hittite, but just like with o and ú, the readings need to be unambiguous across all cuneiform-using languages. So you’ll only ever use bé, pí, and pé in Hittite, never be, pi, or pe.

Except for the V signs, no signs in Hittite ever distinguish o from u—which is another reason why dictionaries alphabetize them together. If it’s important to distinguish them, you’ll learn how to do that in the next section. But as a result, nobody’s really put in the effort to assign readings like do and ko to those signs; in sign lists, you’ll find that bu also has the reading pu, but not bo or po.

In Sumerian and Akkadian, there was a difference between the sounds of b and p, d and t, and g and k. But the Hittite language didn’t have this distinction, and in fact, evidence suggests the Hittite scribes couldn’t even hear it in Akkadian—they constantly mess it up when writing Akkadian words! So in Hittite, the difference between (e.g.) pa and ba just came down to personal and aesthetic preference; generally, ba was used in names and titles like Tabarna (the title of the earliest Hittite kings), and pa was used everywhere else. Soimilarly, the sign gu was only used once across all surviving Hittite texts; everywhere else, ku was used instead. When you’re writing non-Hittite words (like your own name), of course, it’s often a useful distinction to be able to make.

The Akkadian language also had a consonant q separate from k. Hittite didn’t, but the sign qa qa is extremely simple to draw, so Hittite scribes occasionally used it in place of ka, giving it the reading ka₄. It’s still usually transliterated as qa in Hittite texts for simplicity, even though Hittite had no distinct q. Later in these lessons, you’ll come across the word for “you” (zikka) written with ka, ga, and qa interchangeably; if we wanted to be pedantic, we could use the readings ka, kà, and ka₄, to show that they all indicate the same sound, but ka, ga, and qa are easier to type and easier to recognize at a glance.

With these syllables under our belt, we can start spelling some more Hittite words, like pedi “to the place” pé-di. I’d recommend practicing and learning these ones before moving on.

Spacing and punctuation

Unlike many other cuneiform-users, the Hittite scribes were generally very precise in laying out their tablets—to the point that the Hittite style of writing is sometimes known as “typewriter cuneiform”. Words are never broken across lines, sections are marked out with “rulings” across the full length or width of the tablet, and a bit of space is left between words. This might not sound revolutionary by modern standards, but none of these were standard features of ancient writing systems!

The Hittites even had what might be history’s first punctuation mark, the “Glossenkeil” (“gloss wedge”), :. Since the Hittites recorded foreign words in their tablets much more often than other cuneiform-users, they used this sign to mark them, sort of like modern italics.

Apart from the Glossenkeil, cuneiform was unpunctuated. But since practically every Hittite sentence ended with a verb, separating sentences wasn’t much of an issue.

Tangent: Hurrian vowels

Note: This section is about languages other than Hittite, so if you’re only interested in Hittite, feel free to skip ahead.

For the Hurrian language, scribes went a step further in simplifying the system. Like Hittite, the Hurrian language didn’t have any distinction between b/p, d/t, and g/k—so they repurposed these signs to make the vowels less ambiguous instead! In Hurrian, ge is used for ke, ki is used for ki, ku is used for ko, and gu is used for ku. They dispensed with the signs ba, da, ga, di, and tu entirely, but brought back be for pe, which wasn’t used like this in Hittite.

So for the Hurrian language, we end up with this table instead:

This fits the sounds of Hurrian (and Hittite!) significantly better, but tradition is a powerful force, so this simplified inventory was never consistently adopted for Hittite. And since Hurrian is a lot less studied than Hittite, Hurrian scholars just use the Hittite readings for the signs instead of assigning their own.

Exercise 3

How would you write these words in bound transcription?

• e-ku-zi “she drinks”
• pa-ti “to the foot”

How would you write these words in cuneiform?

• apata “therefore”
• aku “she should die”

Memorize these first 18 CV signs, and make sure whatever method you used to learn the V signs scales up well. There are plenty more signs to come, and it’s important to space them out so they don’t get overwhelming.

Lesson 4: CV signs, part 2

Now that you’re used to the idea of CV signs, it’s time to present the rest of them.

The signs cu and cú are used more or less interchangeably. (If you’re wondering what c means, check the section on pronunciation for details.)

Ri and li also have readings re and le.

Ze and he, without the accents, are alternate readings for zi and hi. Zé and hé, with the accents, are distinct signs (shown in the table). I recommend using the accented ones; they’re less ambiguous.

Mi and ni have the less-common alternate readings mé and né. But since there were unambiguous signs for me and ne already, those saw more use.

Plene spellings

Even so, though, many of these signs are ambiguous—when we see the sign le, how do we know if that’s a li or a le? To solve this, Hittite scribes would often use “plene” spellings (Latin for “full”): the V signs indicate every vowel unambiguously, so adding an extra V sign will always make things clear. The word for “don’t”, for example, is always written le-e, which makes it clear that it’s not li.

This is one reason to use a plene spelling, but not the only one—we often find plene spellings with the vowel a, for example, which can never be ambiguous. Since we’re not sure how exactly to interpret them, these spellings are marked with a horizontal lineCalled a “macron”. See lesson 1C for details.

above them in bound transcription: le in bound transcription is le, while le-e is lē. But this may or may not have actually indicated a difference in pronunciation.

Now we can write a few more Hittite words: ha-pí hapi “to the river”, a-ri ari “to a friend”, ka-a-ca kāca “behold”. But we’re going to need a new type of sign to spell longer words. So once you’re comfortable with these signs, move on to the next section to complete the Hittite syllabary.

Tangent: Hyperplene spellings

Note: This is an odd curiosity of the writing system that very rarely comes up. Feel free to skip it.

In a very small number of words, we see what’s known as a hyperplene spelling, where the same V sign (it’s always a, without exception) is written twice in a row. For example, the verb “to be hot”Kloekhorst derives it from PIE *h₁eh₃i-, but I don’t know of any other evidence for a root like that.

has forms like a-a-ri.

In some dialects of Akkadian, there was a convention of using a-a to write the sequence ayya, as in da-a-a-nu for dayyānu “judge”. And since this particular verb also has forms like aitat “it was hot”, it’s possible that there was a yy between the vowels here too. (It couldn’t have been a single y, because aya got contracted to a in the prehistory of Hittite.)

As a result, some authors treat this combination as a new sign with the reading ayya, and transcribe a-a-ri as ayya-ri. But others disagree, and say it just shows there were multiple syllables there, without proving anything about what might have been between them, with the transcription a-a-ri. Given how few words show these hyperplene spellings, it’s possible we’ll never know. But personally, I like the ayya interpretation.

Exercise 4

How would you write these words in bound transcription?

• ka-a “here”
• a-pa-ci-la “personally”
• pu-o-ti “to the salt”

How would you write these words in cuneiform?

• lōri “disgrace”
• ceni “to the model”
• mūlili “the mulili plant”

Take as much time as you need to learn these signs before moving on. There are a lot of them!

Lesson 5: VC signs

With the V and CV signs mastered, you can now write any open syllables: syllables that end with vowels. This works great for a language like Japanese or Māori, where almost all syllables are open.

But Hittite, like English, has a lot of closed syllables, which end with consonants. To handle these, we need a new set of signs: the VC signs, for vowels followed by consonants.

The VC signs distinguish fewer consonants than the CV signs: b/p, d/t, and g/k have to share their signs. This doesn’t cause a problem in Hittite, which never distinguished those consonants anyway—though it is quite annoying in Akkadian, where this difference is important!In Akkadian it’s even worse, since the “emphatic” consonants also aren’t distinguished at the end of a syllable: ak is also ag and aq, it is also id and iṭ, us is also uz and uṣ, and so on. As far as I know, nobody in the Akkadian world ever found a solution for this. But this is actually how iz gets the reading iz: it started as a logogram for “wood”, and the Akkadian word for “wood” is iṣum.

Every single one of those signs has alternate readings—ap is also known as ab, uk is also known as ug, and so on—so that we can write transliterations like at-ta and ad-da, which just look better than ad-ta and at-da. But no matter how we transliterate them, no sign in cuneiform ever distinguishes b/p, d/t, or g/k/q at the end of a syllable. This is a universal rule going back to SumerianIn Sumerian, the “voiceless” consonants were originally aspirated, and stops at the end of a word were unreleased, so the aspiration distinction disappeared.

.

The signs ip, it, ik, im, ir, and iz also have readings ep, et, ek, em, er, and ez.

The sign ah also has readings eh, ih, and uh, but not oh. As above, o gets left out for the simple reason that nobody’s been interested in going through and adding o readings to the sign list. But if you use one of these at the start of a word, it’s a good idea to use a plene spelling to show what vowel you mean!

Like cu and cú, ur and úr are used more-or-less interchangeably.

Writing closed syllables

Equipped with our new list of VC signs, we can now write words like appa ap-pa “back” or eczi ec-zi “is”. But what do we do when a syllable has a consonant at the start and at the end? How can we write per “house” with these signs?

The solution is to combine CV and VC signs with the same vowel. To write per, we use pé-er. Even though the vowel gets written twice, this is not considered a plene spelling; it’s just the standard way of representing these syllables, so we don’t use a macron in bound transcription. If we did want to use a plene spelling—say, to write the word cēr “above”—we would add an additional V sign, ce-e-er.

Equipped with these, we can now write more elaborate Hittite words, like antuhcac an-tu-uh-ca-ac “person”, or haccucari ha-ac-cu-ca-ri “to the queen”.

This system works great for Akkadian, which never allows more than one consonant at the start or end of a syllable. But Hittite, like English, allows more. How can we use this system to write a word like harkzi, “has”?

The usual solution is to use additional VC signs! To write the syllable hark-, we use one CV sign for the consonant at the start, and two VC signs for the consonants at the end, ha-ar-ak. The a vowel of the third sign is considered an empty vowel: there’s no actual vowel there in the Hittite word, it’s just required by the constraints of the writing system.

Sometimes Hittite scribes would repeat the previous vowel in these VC signs, sometimes they’d use the vowel a regardless of what came before, and sometimes they’d use CV signs instead, like ha-ra-ak. For example, linkta “she swore an oath” can be li-in-ik-ta, li-in-ak-ta, or li-in-ka-taAs far as I know, the first of these is the only one attested for this particular form, but we see all of them across various verbs and forms. That’s the problem with having relatively little text to work from!

. Since this problem never came up in Akkadian, there was no standard precedent to follow, and individual scribes (or maybe scribal schools?) came up with their own solutions. But whichever you choose, I recommend always using the VC signs for this instead of the CV ones: if you write hark as ha-ra-ak, how will you write harak?

CV/VC ambiguity

Now that we have both CV and VC signs to work with, sometimes there are multiple ways to write the same word. When writing the word apac “that”, should it be a-pa-ac, or ap-ac?

The answer is a-pa-ac. Even though it takes three signs instead of two, ancient scribes almost always spelled it this way.

Why use the extra sign? The choice actually gives us some insight into how the Hittite language worked! It seems that in Hittite, when there was a single consonant between two vowels, it acted as the start of the second syllable, not the end of the first syllable. In other words, apac was pronounced as a-pac, not as ap-ac.

As a result, whenever there’s an ambiguity like this, you should use the CV signs instead of the VC ones, even if it’s longer. This way, the breaks between syllables line up with the breaks between signs—and it lets us use the VC signs to indicate multiple consonants at the end of a syllable, as we saw above.

This is something the Hittite scribes were generally quite consistent about, and they make sign breaks line up with syllable breaks as much as possible. The only exception to this rule was—rarely—when there was an obvious break between the parts of a wordThat is, the morphemes and/or clitics.

; in this case, they sometimes prioritized those breaks instead. For example, the word ēcer “they were” breaks down into ēc- “be” and -er “they”; it’s usually written e-ce-er, but can also (rarely) be written e-ec-er.

Exercise 5

Write the following words in cuneiform.

• harcan “head”
• icpanza “night”
• canhti “you search”

Write the following words in bound transcription.

• ha-al-ki-ic “grain”
• ec-ha-ar “blood”
• ka-ar-ap-zi “she lifts”

Start memorizing the VC signs. This is the largest batch you need to memorize in this course; it’s all downhill from here.

Lesson 6: The last few signs

This set of signs is almost sufficient to represent every syllable in Hittite. But there are two consonants missing: y and w. These sounds are rare to nonexistent in Akkadian, so when the Hittites adapted the writing system, they didn’t have much material to work with here.

Y

For y, there’s a special sign ya, which is just i and a smashed together: ya. There are no signs for y in any other combination—that is, no signs for ye, yu, ay, and so on. Yi probably wasn’t possible at all.

This is partially because these combinations were much rarer than ya. But when they did have to write a word like yugan “yoke”, they would use the i sign: i-ú-ga-an. Even in the case of ya, it was occasionally written with an extra i for good measure: the serpent Illuyanka, who you’ve seen in some of the examples, is alternately called il-lu-ya-an-ka and il-lu-i-ya-an-ka. As a result, some scholars question whether there was a difference between y and iy at all.

W

W is rare to nonexistent in Akkadian, depending on the dialect, but even more common than y in Hittite. So the Hittites had to get creative for this one.

Some older dialects of Akkadian have a sign for wa, wa, which the Hittites used extensively. This sign is better-known to Akkadian scholars as pi, which is how the later, w-less dialects of Akkadian used it; this is the reason why Hittite only uses pí for that syllable. (Remember, the sign names have to be unambiguous across all languages that use cuneiform, not just Hittite!)

In addition, the Hittites repurposed the logogram ŊECTIN ŊECTIN, which means “wine”. Their word for “wine” was wiyanac, so they used this sign for the syllable wi, giving it the reading wi₅.

For w in front of any other vowel, they’d use the sign ú (or less often o). The word welluc “meadow”, for example, would be written ú-el-lu-uc. Wu and wo might not have been possible at all. Since wi₅ took a while to catch on, you’ll also not infrequently find ú-i for wi. Just like with iy, we sometimes see uw, and there may not be any difference between that and normal w.

cp-

You’ve seen now how to write multiple consonants at the end of a syllable. But what about multiple consonants at the beginning?

Well, the prevailing view is that it just didn’t happen in Hittite. There are a lot of words that start with the combinations icp-, ict-, and ick-, always written with an i rather than any other vowel, which suggests that Hittite did something similar to Spanish: if there were multiple consonants at the start of a word, add a vowel in front to break them up. (That’s why we see Spanish escuela from Latin schola “school”, espada from Latin spatha “sword”, and so on.)

However, there are a few times when these words are written with ci- instead of ic-: the word icpanti “she sacrifices” is usually ic-pa-an-ti, but is sometimes ci-pa-an-ti instead. Some scholars take this as evidence that there wasn’t really an i there at all, and the word should really be read as cpantiIf this is the case, though, it’s not clear why they’d consistently use an iX sign, because multiple consonants at the end of a syllable generally use aX or repeat the vowel.

.

When writing names with multiple consonants at the beginning, I recommend inserting an i at the beginning; it’s likely that the ancient Hittites would pronounce an extra vowel at the start of “Steven” just like modern Spanish-speakers do. If your name really does start with a sequence like isk-, you can use a plene spelling to make it clear.

A brief example

Using these signs, we can write even more Hittite and non-Hittite names. For example, Wiluca wi₅-lu-ca was a city-state in ancient Anatolia. One king of Wiluca, Alakcandu a-la-ak-ca-an-du, made a treaty with the Hittite emperor Murcili II mu-ur-ci-li, invoking the patron god of the city, Āppaliona a-ap-pa-li-o-na. In the treaty, Murcili claimed that Wiluca and the Hittites had been on good terms for centuries. Later in his reign, though, Murcili needed to reassert control over Wiluca by force—the exact circumstances aren’t clear, but one theory is that Alakcandu was overthrown by a faction more sympathetic to the foreign state of Āhhiyawa a-ah-hi-ya-wa.

It’s hard to be sure from the very fragmentary documents we have, but it’s possible that we’re seeing here a contemporary account of the Trojan War! Wiluca could be Wilion (another Greek name for Troy, hence the “Iliad”), Āhhiyawa could have been the Achaeans (that is, the Mycenaean Greeks), Alakcandu could have been Alexander of Troy (better known as Paris), and Āppaliona could have been Apollo, the divine defender of the Trojans.

If you look for these names online, you’ll more often see them transcribed as Wilusa, Alaksandu, Mursili, Appaliuna, and Ahhiyawa. In other words, c is usually transcribed as s. This conveys the pronunciation better when talking about the Hittite language, but it’s very important to write c (or the fancier š) instead of s when you’re talking about cuneiform, for reasons you’ll see in a later lesson. Just remember to change it to an s when you’re searching.

Tangent: F and V

Note: This section is not important for writing Hittite. If you’re only interested in Hittite, feel free to skip down to the next one.

And that’s all the sounds of Hittite! But unlike Sumerian and Akkadian scribes, Hittite scribes also cared a lot about representing the sounds of other languages as accurately as they could. The Hittites called themselves “The People of a Thousand Gods”, and they tried to adopt the deities of conquered peoples into their pantheon whenever possible. And with those adopted gods came adopted prayers and rituals—if you want a Hurrian god to respond, then obviously (at least to a Hittite priest) you need to pray in Hurrian. Otherwise, how will the god understand you?

They weren’t always successful in recording these sounds. For example, they couldn’t tell apart the voiced and voiceless sounds of Akkadian, so incantations in Akkadian are filled with confusion between t and d, p and b, and so onThese incantations are marked Babilili—“in the language of Babylon”—but they usually wrote it pa-bi-li-li.

. But they did recognize that Hurrian and Hattic had a sound that didn’t line up with any of their cuneiform signs, and invented some new signs to represent it.

These new signs consist of wa with the vowel signs a, e, i, o, ú written under the horizontal stroke: wa+a we+e wi+i wu+u wu+ú. (Scholars disagree on whether o and ú represented different vowels in Hittite, but they definitely did in Hurrian.) Since these signs were never used to write Hittite words, the sound must have been something different from the Hittite w; the leading theory is that it was a labial fricative, something like the English /f/ or /v/. Since we’re still not sure quite what sound this was, these signs are most often transcribed as wa_a, we_e, and so on, or as wa₄, wé, wí, wú, wù—but I also assign them the readings fa, fe, fi, fo, and fu. Even if the sound wasn’t the same as the English “f”, after all, we know that it wasn’t a Hittite /w/, and it makes things more readable to assign it its own letter instead of repurposing w or adding even more subscripts.

So while these signs aren’t needed for actual Hittite words, they can be useful for writing foreign names (like yours!). Hurrian seems to have had the same length distinction as HittiteThat is, distinguishing between tt and t rather than between t and d.

, so to make a long f, they would use the p signs: affa would be af-fa, uffe uf-fe, and so on. Hurrian scholars generally give these signs the readings áw, éw, íw, úw (for transcriptions like awwa and uwwe), since Hurrian didn’t have the Hittite w. But I prefer the readings af, ef, if, uf, for the same reason that we don’t write the Hittite c as s—it’s cleaner not to change transcription systems in the middle of a text.

For example, the name of one of the (many!) Hittite sun gods was borrowed from Hattic: fu-ru-un-ce-mu. This is usually transcribed as Wuruncemu, but I prefer the transcription Furuncemu.

Four additional signs combine wa with ap, ip, pí, and pu: wa+ap wi+ip wi+pí wu+pu. It’s not clear what sounds these were supposed to represent, so they don’t have standard readings assigned, but it probably involved f in some way. They could be something like *áf, *íf, *fí, and *fú, but it’s not clear why the scribes would invent both fi and wi+pí, or fu and wu+pu, and they were perfectly happy to use ap for af (and so on). So what exactly these signs meant is an open question in Hittitology, and I haven’t attempted to give them any fancy names myself.

Exercise 6

Learn the Y and W signs (and the F signs, if you like), and start practicing all the V, CV, and VC signs together. You won’t need to memorize any more signs in this course—the rest can all be looked up as you need them—but it’s vital to have at least a basic memory of all the V, CV, and VC signs, since they make up the vast majority of any Hittite text.

Now that you have them, though, it’s time to start working with some actual Hittite documents! Remember the story of Illuyanka? Make a bound transcription for these lines, where the goddess Inara has made her proposition to the human Hopaciya.

ki-ic-ca-an hu-o-pa-ci-ya-ac i-na-ri
Hopaciya spoke thus to Inara:

ma-a-wa ka-at-ti-ti ce-ec-mi nu-wa ú-wa-mi
“If I can sleep with you, I will come.”

Convert these sentences into cuneiform. The equals signs mark the boundaries between clitics, where multiple units are combined into a single word—like how “men’s” is a single word in English, but it’s made up of “men” and “’s”. You can ignore them when writing your cuneiform: in cuneiform, n=ac and nac are written exactly the same.

“Kardiac=tac iyami.”
“I will fulfill your wish.”

N=ac katti=ci cecta.
So she slept with him.

I’ve modified the cuneiform for these examples slightly, to avoid using any signs from later lessonsFor example, there really should be some determiners.

, but the Hittite words are the same. You’ll find the full, unadulterated Hittite text in the final test.

Lesson 7: CVC signs

In addition to the V, CV, and VC signs, Hittite had a limited selection of CVC signs, representing a vowel with a consonant both before and after itIn the original Sumerian cuneiform, most signs were actually CVC signs: an was originally ʔan, ba was originally baʔ, and a was originally ʔay. But even within Sumerian these quickly got simplified to the readings we’re now familiar with.

. Unlike the others, these signs weren’t strictly necessary: the word wātar “water” could be written wa-a-tar, with a CVC sign tar, but it could just as well be written wa-a-ta-ar, with a CV sign ta and a VC sign ar. It’s a tradeoff between efficiency and complexity: tar is much shorter than ta-ar, but it also means a new group of signs to learn.

If there was a unique sign for every CVC combination in the language, that would require at least 520 signs—more than there were in the entire Hittite writing system! So many of these signs had multiple readings, and many possible CVC combinations had no signs for them at all. Scribes tended to use these signs however they pleased to make words shorter, with some scribes using lots and others using none at all. You can learn them if you like, or stick to CV-VC pairs and skip to the next section. Just don’t be surprised if you come across them in a text.

No cuneiform signs ever distinguished b/p, d/t, or g/k/q at the end of the syllable, so every sign like hap, without exception, also has a reading hab. Since Hittite didn’t distinguish those consonants in general, most signs like kam also have readings like gam; I haven’t listed those here to save space, but you can find them in a dictionary.

Two special cases

However, there are two words in Hittite that are exclusively written with CVC signs. It’s important to know these ones even if you ignore all the rest, because these meanings are almost never written with CV-VC pairs.

The “emphatic particle” =pat can be attached to the end of any word to emphasize it, sort of like an underline (or rather an underline that can be read out loud). This word is always written with the CVC sign pát rather than pa-at. In Akkadian, that sign also has the reading be, but in Hittite, bé (bé) is used for that instead. There’s historically been some confusion over how this sign should be read—since the particle is always written this way, how do we know which way it’s supposed to be pronounced? So you’ll find older works transcribing it as be or pít; nowadays, though, seeing it used as a CVC sign in other words has led to a consensus on pát. For example, the goddess Hepat is sometimes hé-pát and sometimes hé-ba-atRemember, there’s no difference between b and p in either Hurrian or Hittite. It would be more accurate to transcribe the longer version as hé-pá-at, but ba is easier to type and easier to read, so it persists.

.

The “sentence particle” =kan is attached to the first word of a sentence to alter the meaning of the verb in hard-to-predict ways. It’s one of the most common words in Hittite, especially in later eras where it started replacing all the other “sentence particles”, but also one of the hardest to translate. Like =pat, it’s always written with the CVC sign kán rather than ka-an, but unlike =pat, there’s never been much confusion over how to pronounce this one.

Apart from these, I’m not aware of any word in Hittite that’s exclusively written with CVC signs. You’re welcome to stick to V, CV, and VC signs for all the rest; while wātar “water” is usually written wa-a-tar, it’s perfectly valid to write it wa-a-ta-ar if you’re not pressed for space.

Exercise 7

Write the following words in two different ways each.

• marri “very”
• kiccan “thus”
• aggātar “death, ruin”

You don’t need to memorize the CVC signs; there are plenty of them that even I don’t have memorized, and just look up when I need them. But if you want to work with actual Hittite texts, make sure you learn pát and kán, since they’re by far the most common.

Lesson 8: Logograms

The approximately 100 signs you’ve learned in the V, CV, and VC categories are enough to write the pronunciation of any word in Hittite. It just might take a lot of them! To write antuwahhac “person”, for example, it takes six: an-tu-wa-ah-ha-ac.

But back in Sumerian times, cuneiform started as a logographic writing system. Like in Chinese, every sign originally represented a specific word, before they were repurposed to write specific sounds. So when an Akkadian or Hittite scribe wanted to shorten a word, they could bring back one of the old Sumerian logograms (“word signs”). Instead of spelling out antuwahhac with those six phonetic signs, you could write the single logogram that means “person”, LÚ. Much shorter!

There are about 300 different logograms that we find in Hittite texts, though most of them are extremely rare. It’s likely that scribes looked up the rarer ones when they needed them rather than memorizing them all, and that there were more logograms in their dictionaries that weren’t used in any surviving texts.

I encourage modern students to do the same: when you come across a sign you don’t recognize, learn to look it up rather than trying to memorize them all. But a handful of logograms are extremely common—common enough, in some cases, that we’re not sure how the Hittite word was pronounced, since it was never written phonetically! A sample of those are included here.

Since we don’t always know how these logograms were pronounced in Hittite, we transcribe them with their original Sumerian names, in ALL CAPS to show that they’re not actually read that way. Many signs can be used either phonetically or logographically, so it’s important to pay attention to the difference: the sign A might be the phonetic sign a, or the logogram A, meaning “water”. The Sumerian word for “water” was pronounced a, which is why this sign was used for this pronunciation in the first place, but in Hittite, it needs to be read as wātar.

(Instead of “logograms”, these signs are sometimes called “Sumerograms”, because they’re taken from the Sumerian language.)

Sumerian phonetics

Since logograms are transliterated with their Sumerian names, we do need to know a little bit about Sumerian phonetics. For the most part, we see exactly the same consonants in Sumerian as in Hittite, which isn’t a coincidence: since the cuneiform writing system was originally designed for Sumerian, the Sumerian inventory was the raw material the Hittite scribes had to work withThe story is actually a bit more complicated than that: additional signs were adapted for the q, ṭ, ṣ, and ʔ sounds of Akkadian, for example, which the Hittites could have used. But overall, they didn’t.

.

But Sumerian also had at least three more sounds that Hittite lacked. First and foremost, it had an s sound distinct from c. While Hittite never had separate signs si and ci, Sumerian very much did—and both of those signs are used as logograms in Hittite! Since SI SI “horn” and CI CI “eye” are different signs, we need to keep their transliterations distinct, and this is why we’ve been using c in our Hittite transcriptions all this time. Having SI refer to SI but si refer to ci would just be too confusing. It’s fine to use the letter “s” when talking about the Hittite language in general, or Hittite names and words in a non-cuneiform context, but if you’re doing anything with cuneiform transcriptions, it’s vitally important to keep your c (or š) separate from your s.

The second sound is the “ng” at the end of the English word “sing”, and is transcribed as Ŋ/ŋ (a cross between “n” and “g”) in these lessons. For ease of typing, you can also represent it as j, which looks vaguely similar and wasn’t being used for anything else; you’ll also find it written as ĝ or g̃ in some dictionaries. Unlike in English, this sound can appear at the beginning of a word as well as the middle or end: ŊIC ŊIC “wood”, DIŊIR DIŊIR “god”, SAŊ SAŊ “head”.

However, most of our knowledge of these Sumerian pronunciations comes from bilingual dictionaries written by Akkadian-speaking scribes. And just like how Hittite-speaking scribes struggled to tell the difference between t and d, Akkadian-speaking scribes struggled with the distinction between ŋ and g, writing it as ng between vowels and g anywhere else. As a result, you’ll usually find those signs written as GIC, DINGIR, and SAG instead, unless the author is really interested in accurate Sumerian pronunciationWhich, well, I am.

. For Hittite purposes, it’s not a problem to spell them that way instead.

The last additional sound is usually transcribed as dr, since that’s how the Akkadian-speaking scribes recorded it. We’re not entirely sure how it was pronouncedThe leading theory is an aspirated version of z, something like /ts^h/.

, but Sumerian scholars believe it turned into a r in some dialects and a d in some others, so dr is a convenient approximation, as in GUDR GUDR “cow”. Specialized books on Sumerian sometimes write it as ř, as in GUŘ, but I’ve never seen this done for Akkadian or Hittite texts, so I’m sticking to dr.

It’s possible that Sumerian had a few additional soundsJagersma suggests IPA /j/, /h/, and /ʔ/.

, but none of them were represented in the bilingual dictionaries, so we don’t use them in transliteration.

Common logograms

This table gives some of the most common logograms in Hittite. There are hundreds more, though; an important part of reading cuneiform is being able to look up new logograms in a dictionary when you need them.

For many of these we know the Hittite pronunciations. The Hittite word for “god” is ciuc, for example, “king” is haccuc, and “land” is utnē. But for many we don’t, because the word is only ever written with a logogram! We can only guess at the Hittite words for “woman” and “son”, for example.

UD is by convention transcribed UD when it means “day” and UTU when it means “sun”It also has the value u₄, for kind of an odd reason. In Akkadian, the word for “day” is umum, and this is usually written with a phonetic complement as UD-mum. As a result, to make their transcriptions cleaner, some Assyriologists like to write this as U₄-mu-um instead of UD-mu-um. I’m not sure why this is done specifically for “day”; the word ilum “god” is usually written DIŊIR-lum, but nobody is giving DIŊIR the reading I_x.

, but this ends up being more of a guideline than a hard rule. Many people split the difference and go with UT, which is how it’s used syllabically even if it doesn’t quite match either of the Sumerian words; just make sure you don’t mix it up with UDU (UDU) or URU (URU)!

Inflection

This system hits a snag when it comes to inflection. In Hittite, words are modified in various ways to indicate their role in a sentence: the word for “person” is antuwahhac if it’s the subject of the sentence, antuwahhan if it’s the object, antuwahhi if it’s the recipient, antuwahhec if there’s more than one person, and so on. When you write the Sumerogram LÚ, how do you know which of these it means?

To some extent, they could work around this problem with more logograms. Sumerian didn’t usually mark if nouns were singular or pluralThough they’d sometimes repeat a logogram, so we occasionally see KUR.KUR for “lands” and such in Hittite texts.

, but the word MEC meant “they are”, so it was only used with plural nouns. So, Hittite scribes borrowed it as a plural marker, using LÚ to mean “person” and LÚ.MEC to mean “people”. Originally there was a convention to use MEC (“they are”) for people and HI.A (“mixed”) for objects, but over time, both of these were applied freely to all sorts of nouns. In transliteration, multiple signs making up a single logogram are separated by dots, so this would be transliterated as LÚ.MEC or LÚ.HI.A.

Unfortunately, Sumerian didn’t have any convenient words that could be repurposed for “subject of a sentence” and so on. The easiest solution, then, was to write the last sign or two of the Hittite word after the logogram—antuwahhac is LÚ-ac, antuwahhan is LÚ-an, and antuwahhi is LÚ-hi.

The same went for verbs: icpanthi “sacrifice” could be written with the five signs ic-pa-an-at-hi or the single sign BAL. But to distinguish between icpanthi “I sacrifice”, icpanti “she sacrifices”, icpantanzi “they sacrifice”, and so on, these could be written BAL-hi, BAL-ti, BAL-an-zi. If you’re used to Japanese, this is very similar to how Japanese-speakers combine kana and kanji!

Phonetic complements

Writing the last sign of a word like this is called a phonetic complement, and it’s useful for more than just inflection. For example, Sumerian only had a single word for “back”, EŊIR. But Hittite has two: appa for “backward” and appan for “behind”. Phonetic complements can be used to tell them apart: EŊIR-pa means appa, and EŊIR-an means appan!

It also offers us a little bit of information about the Hittite word, even when we don’t know the whole thing. We’re not sure quite how “woman” and “son” were pronounced, for example, but they’re sometimes written with complements like MUNUS-za and DUMU-ac, so we know they end in -za and -ac. This can be a big help when we’re trying to figure out the words! With a whole lot of this sort of detective work, for example, Melchert figured out that the word for “son” had to start with a vowel and end with -lac, and proposed that the mysterious term ayawalac found in one particular letter is actually the missing word for “son”. Similarly, Güterbock analyzes the various different phonetic complements used with MUNUS and compares them against words in other languages to propose the reading *kuwanzaThe star in front means this word is never actually attested anywhere, just reconstructed by linguists.

.

Numbers

One of the most popular uses for logograms, across all sorts of languages and cultures, is for numbers. It’s just so much more convenient to write “69,105” than to spell out “sixty-nine thousand, one hundred five”, especially when you’re filling out tables or doing accounting. Part of why the cuneiform writing system developed in the first place was for accounting and tax purposes, so it’s no surprise that they’ve got a well-developed number system as well!

This system went through a lot of different changes over the millennia it was used, but this is how the Hittite version works.

Number signs in Hittite are made up of two components: vertical strokes, and hook/diagonal strokes. A vertical stroke represents 1, and a hook or diagonal represents 10.

These symbols can be combined to write larger numbers. “29”, for example, would be 20-9.

This system works as you’d expect up to 59. But 60 is not written with six hooks—it’s written with a single vertical again!

Originally (back in the Sumerian days), 1 was a narrow stroke and 60 was a wide stroke, written with two different round styli10 and 600 were then written by stamping the end of the round styli into the clay to make differently-sized circular impressions.

. But when scribes started using a single rectangular stylus to write, this turned into the world’s first place value system! Just like how the digit “1” for us can mean 1, 10, 100, or 1,000, depending where it appears in the number, a single vertical stroke can mean 1, 60, or even 3,600. So the number 86 would be written as 80-6: one vertical for 60, two hooks for 20, and six verticals for 6.

Since the Sumerians had no zero, the sign 60 was generally not used on its own in Hittite—it was just too ambiguous. To write the number 60 in isolation, the Hittites borrowed the Akkadian word ŠU-ŠI.

Like all logograms, these signs do have fancy Sumerian names (DIC for 1, MIN for 2, EC₅ for 3…), but it’s more common just to write their valuesSome older sources like to write these values in Roman numerals, like XVII instead of 16. The meaning is the same.

. Since some of them look very similar to other signs—for example, 10 (10) looks identical to o (o), and 90 (90) looks very similar to MEC (MEC)—the Hittites also had a special logogram, KAM, meaning “number”. This makes it clearIn earlier styles of writing, MEC and 90 were truly identical. It’s only in later Neo-Hittite script that MEC started to be written with horizontals instead of hooks.

that UD.MEC is UD.MEC “days (plural)”, while UD.90.KAM is UD.90.KAM “ninety days”.

While the Hittites definitely had words for the numbers, these logograms were so convenient that their pronunciation is almost never written out. We only know how 3 and 7 were pronounced because they were used to spell other words: the drinks teriyallac and siptamiyac are sometimes written as 3-yallac and 7-iyac, suggesting that those two are pronounced teri- and siptam-. Scholars have made guesses at the others, but they’re approximate at best—some people think 1 was anzac, for example, while others say ciac.

Tangent: Larger numbers

Note: Like the other tangents, this section is entirely optional. Feel free to skip ahead if you’re not interested in mathematics.

In Babylonian cuneiform, you could use the place value system for numbers as large or as small as you want—the “1” sign could also mean 1/60 or 1/3600, extending the place value system to the right just like we do with our modern decimal point. But the Hittite scribes had no need for this kind of fancy mathematics, and their own number words were based on tens, not sixties. So they adopted separate logograms for “hundred”, “thousand”, “ten thousand”, and “hundred thousand”I’m simplifying somewhat here: we don’t actually see the single sign LIM used for “thousand” on any Hittite tablet. Instead, they spelled out the Akkadogram LI-IM. But the progression is clear enough, we have very few attestations of numbers this large, and the LIM sign is frequently used for that pronunciation in other Akkadograms, so I don’t think this it’s unreasonable to extrapolate that LIM on its own could also be used this way. Similarly, the sign for “hundred thousand” is reconstructed only, but we can reasonably speculate that it was originally a single sign for the reasons given below.

.

100 is just the sign me, short for meʔatu, the Akkadian word for “hundred”. 1000 is a straightforward combination of 10 and 100, which thus took on the reading lim in Akkadian; līmu is the Akkadian word for “clan” or “tribe”, and the title for “leader of a clan” was repurposed to mean specifically “commander of a thousand soldiers”. The fact that it looks the same as ci (the syllable ci or a logogram for “eye”) is sheer coincidence.

10000, then, is 1000 with some extra strokes added; in Akkadian, this is called a gunû sign, from the Sumerian for “decorated”. It’s once again coincidence that it looks the same as the logogram SIG₇ “yellow-green”. But we don’t know how this one was pronounced in Sumerian, Akkadian, or HittiteThough phonetic complements suggest the Hittite word ended with an -a!

, so 10000 or SIG₇ is as good a name as we’re going to get.

There’s no sign for 100,000 in Akkadian or Sumerian. It’s just not a number that ever comes up in everyday life; for advanced mathematics, the Babylonians could use their place-value system, and to express a vague big numberLike how an English-speaker might say “it feels like a million degrees out”.

, they’d use 3,600 (that is, 60 times 60—it’s not a coincidence that this is the number of seconds in an hour!). But there’s exactly one Hittite text where this number is expressed, and it was probably expressed with the sign 100000—that is, 10000 with even more extra strokes.

I say “probably” because this text was composed in the Old Hittite period, but the only version that survives is a later copy, and the later scribe clearly didn’t recognize this sign. He replaced it with the sequence GACAN-TI. The logogram GACAN means “lady”, and the Akkadian word for “lady” is bēlti, so the scribe presumably read this as a logogram with an Akkadian phonetic complement. But in context, the word “lady” makes no sense whatsoever.

The relevant line readsWell, almost. As mentioned above, I modified the Akkadogram LI-IM to LIM here for simplicity, since Akkadograms aren’t covered until lesson 10. The same goes for the other texts in this section, which similarly used LI-IM in the original.

9-an 1000 9-an 10000-an 9-an 100000 MU.HI.A-uc: “nine thousands, nine ten-thousands, nine hundred-thousands years [may the gods grant to the king and queen].” So this seems to be how the Hittites wrote out larger numbers: a single digit (1 through 9) followed by a logogram for the order of magnitude. These signs don’t seem to be used alone; for something like “one hundred” they write 1 1000 instead of just 1000. And they aren’t always written from smallest to largest; we also have lines like 1 10000 5 1000 5 100 NAM.RA.MEC “one ten-thousand, five thousand, five hundred captives”.

These numbers are extremely rare; generally, if a Hittite scribe wanted to express a vague big number, they’d use “1,000” (for example, the Hittite pantheon is known as 1 1000 DIŊIR.MEC, “The Thousand Gods”), and there’s very little other need for numbers higher than 99. As a result, most people name these logograms ME, LIM, SIG₇, and GAŠAN+TI, based on the signs they look like. But I’ve tentatively assigned them the names “100”, “1000”, “10000”, and “100000”, for easier reading. Thus, we can transcribe the line above as “9-an 1000 9-an 10000-an 9-an 100000 MU.HI.A-uc”, without having to resort to the logogram names for “yellow-green” and “lady”.

As a final note on this topic, the number “1,000” seems to have had special importance to the Hattic people who predated the Hittites. The Hattic word for “thousand” is written as fa-ar (remember the fa sign?), and it appears in the Hattic word for “pantheon”, ta-fa-ar-wa-ac-ha-ap. It may also be in the middle of the Old Hittite royal title Tabarna, which could have meant something like “the ruler of a thousand” (that is, “everyone”).

Exercise 8

You don’t have to memorize all the logograms in the Hittite language—that would be a monumental task! If you do more work with Hittite, you’ll start to recognize the most common ones, but for all the rest, even experts in cuneiform have to consult a dictionary from time to time. You’ll find some options in the “further reading” section.

Here are five Hittite words, written both phonetically and logographically, but in scrambled order. Transcribe the pronunciation of each word, and match the pronunciations to their meanings.

• KUR.KUR-e
• ci-uc
• pi-ir
• DIŊIR-uc
• wi₅-ya-an-ac
• ci-wa-at-ta
• É-ir
• ut-ne
• ŊECTIN-ac
• UD.3.KAM-at-ta

Lesson 9: Determiners

Unfortunately, Sumerian logograms were extremely ambiguous. The pa sign started as a picture of a branch (pa in Sumerian) with some twigs coming off it. But it was also used to write the Sumerian words ŋidru “staff” (since a staff looks pretty similar to a branch) and ugula “overseer” (it’s easier to draw a staff than a person who wields one). Sometimes the difference between a staff and an overseer is important—so how do you indicate which is which?

The solution was to add a determinerThe linguistics term is semagrams, “meaning signs”. But this is practically never used in Assyriology.

: an extra sign indicating what general sort of thing you’re talking about. An overseer is a person, so it’s written LÚ^UGULA; a staff is made of wood, so it’s written ŊIC^ŊIDRU. In transliteration, determiners are generally written in superscript: ^LÚUGULA, ^ŊICŊIDRU. When typing, you can use a caret instead: LÚ^UGULA, ŊIC^ŊIDRU.

These determiners proved useful enough that they started being used with Akkadian and Hittite words, too. They can be a useful clue to what a word means, and there are various Hittite words whose meanings are only known through their determiners: we aren’t sure what exactly an urāc is, but it’s written with the determiner DUG, so it must be some kind of “vessel”.

Some of the most common determiners are listed here, but a lot of logograms can be used as determiners in certain circumstances. It’s not always clear if a sign should be read as a logogram or a determinerThis is especially a problem when there are multiple determiners on a word. The Hittite empire was known as KUR-URU-ha-at-ti, for example, “the land of the city of Hatti”; did the Hittites actually pronounce a word for “land” there, or “city”, or both, or neither?

: perhaps LÚ.ŊIDRU should actually be read as LÚ.ŊIDRU “person of the staff” instead of ^LÚUGULA “overseer (who is a person)”, with LÚ being part of the logogram? The difference isn’t especially important: either way, the Hittites pronounced it something like maniyahhatallacWe know the verb for “manage, oversee” was maniyah-, so it was probably some derivative of that, even if we’re not sure of the particulars.

, and wrote it with the signs LÚ-UGULA.

Common determiners

Most determiners come before the word, but the ones with a dash in front of them come after: ha-ra-an^MUCEN “eagle”.

The DIC determiner can be abbreviated to 1 or m in transliteration, the MUNUS determiner to f, and the DIŊIR determiner to d; these three specifically can be used in front of proper names, to indicate what sort of person they’re talking about, so it’s useful to abbreviate them as short as possible. In some eras, m and f were used for male and female names respectively, so they’re transliterated as m and f; in other eras, 1 was used for names of any gender, and then the transliteration 1 is betterThe former convention (m for men, f for women) is more common in Hittite, but we find both used. In Akkadian, we sometimes even find 1 for men and 1-f for women. The 1 transcription, of course, comes from it being the logogram for the number one.

. Some older publications even transliterate it as I—that is, the Roman numeral for 1—but having I mean 1 and i mean i seems unnecessarily perverse. I highly recommend using 1 instead (ideally in a font that makes them easy to tell apart).

In Hittite, LÚ was used for male professions (“types of men”), and MUNUS for female professions (“types of women”). Hittite society was not especially egalitarian, and there were practically no professions that didn’t have gender associations. In Akkadian, on the other hand, LÚ is used for any sort of profession (“types of person”). When writing modern, less-inherently-gendered words, I recommend using 1 for any personal name and LÚ for any type of person.

Since d indicated that something was a divine name, that divine name could be cut pretty short. For example, UTU can be used as a logogram for “sun” or “day”, so d^UTU can stand for the name of any sun god. But given that the Hittites did everything they could to adopt foreign deities into their pantheon, this creates a lot of ambiguity for modern scholars: we know the Hittites worshipped at least three different sun gods, but it’s hard to even make a guess at their namesAt the very least, there’s “the Sun Goddess of the City of Arinna”, “the Sun God of Heaven”, and “the Sun Goddess of the Earth”, who probably included some combination of Arinniti (“of Arinna”), Hittite Istanu, Hattic Furunsemu, Mesopotamian Shamash, Indo-European Tiwaz, and Hurrian Simige, but there’s a lot of disagreement on which name should be associated with which title.

! This is why English translations tend to refer to “the Sun God” and “the Storm God” rather than using specific names—that’s literally what the cuneiform signs say. We are perhaps fortunate that the Hittites didn’t go as far as the Mesopotamians did in referring to the gods by their holy numbers.

Exercise 9

Like the logograms, you don’t need to commit all the determiners to memory. Learn at least 1/m, f, and d, so that you can recognize proper names, and you’ll pick up the rest as you go along.

We don’t know the exact meanings of any of these words in Hittite. Take your best guess at what they mean.

• ŊIC^e-ya-an
• DUG^ka-ap-pi
• LÚ^pa-ti-li-ic
• TÚG^te-er-ma-az
• la-ak-ka-ar-wa-an^SAR
• NINDA^ha-ar-ci-ic
• É^ci-na-ap-ci
• ú-ra-ya-an-ni^MUCEN

Lesson 10: Akkadograms

While logograms are used to make the writing more concise, the last type of sign does the opposite—these are generally used to make it more verbose!

In Hittite times, Akkadian was the most common language used all across the ancient Near East. Sumerian was a dead language only studied by priests and scribes, Egyptian wasn’t much used outside Egypt, and Aramaic hadn’t become prominent yet; Akkadian was the language of international diplomacy and the cultured elite. And after all, the Hittites had adapted the cuneiform writing system from Akkadian-speaking scribes.

As a result, inserting Akkadian words here and there in your Hittite writing was a good way to demonstrate how well-educated you were! These weren’t logograms; they were spelled out phonetically, to the best of the scribe’s ability. They were just words written in a different language, which you were supposed to substitute for their Hittite equivalents when reading. It’s a bit like writing i.e. instead of “that is” or e.g. instead of “for example” in English—except that for full effect, you’d write out the full Latin phrases id est and exempli gratia, taking up just as much space as the English would!

The linguistic term for these is “heterograms” (literally “different words”), but they’re more often called “Akkadograms”, since the language they’re written in is always Akkadian. These would never be used for more than a word or two at a time; while most Hittite scribes knew a few Akkadian words, few could actually speak it fluentlyWe know this based on how many errors they made!

.

In transliterations and bound transcriptions, Akkadograms are written in ITALIC CAPITALS.

Akkadian phonetics

Since Akkadograms (unlike Sumerograms) are actually spelled out phonetically, it’s important to talk a little bit about the sounds of Akkadian. The Akkadian spoken during Hittite times had five additional sounds that Hittite didn’t: s, q, ṭ, ṣ, and ʔ.

ʔ is the easiest to dispense with. It was pronounced as a glottal stop—the pause in the middle of “uh-oh”, or the Cockney pronunciation of “bottle”. But it was never written consistently even in Akkadian, and Hittite scribes didn’t write it at all. For Akkadogram purposes, it can be ignored completely.

S was a sound that was different from c, though how exactly it was different varied by dialect. Most modern scholars pronounce c as /ʃ/ (the “sh” in “ship”) and s as /s/ (the “s” in “sip”), which is how it was pronounced in Assyrian dialects; in Babylonian dialects, though, it was the other way around. The Hittite scribes generally did a okay job of transcribing this one, though they did often confuse it with z.

Q, ṭ, and ṣ are known as the “emphatic” consonants, and were pronounced like k, t, and s with some additional elementPhonetically, they were probably ejectives. Akkadian was a Semitic language, and these are cognate with the emphatic consonants of Hebrew, Arabic, etc.

. The Hittites were much worse at distinguishing these ones, and frequently mixed them up with k/g, t/d, and s/z, respectively. To make matters even worse, Sumerian didn’t have any of these sounds, so Akkadian scribes didn’t inherit any signs to write them with! As a result, even in Akkadian, these are often written with the signs for other sounds, and modern scholars generally don’t even try to pronounce them distinctly.

The signs used for these sounds are given below.

Note that Akkadian only had four vowelsIt also had three different levels of vowel length: a is never written plene, ā is sometimes written plene, and â is usually written plene. For Hittite purposes, this doesn’t matter, but don’t be surprised if you see a circumflex on an Akkadian vowel.

, and only distinguished three of those reliably.

Ṣi is used for zé in Hittite words. This is a Hittite innovation: in Akkadian, ṣi is used specifically for ṣ and zi specifically for z, but the Hittite scribes couldn’t hear the difference, and repurposed zé to make the vowels less ambiguous instead.

I’ve included this block of signs because I find them interesting, but it’s important to note that the Hittites never used them very consistently! Most Hittite scribes were not very good at Akkadian, and weren’t entirely sure where the special “emphatic” signs were supposed to go.

As a result, they’d very often use k/g, t/d, and z signs instead. Thanks to the HittitesWell, not just the Hittites; I’m simplifying a bit here. Since Sumerian didn’t have a q sound, Akkadian scribes didn’t inherit any signs for it, and until they came up with their own they just used the k and g signs.

, ga also has a reading qá and ka has a reading qà, for example—the scribes just used the signs they were familiar with, since they couldn’t hear any difference between q, k, and g. And conversely, qa also has a reading ka₄, since to the Hittites, it was just an easier-to-write symbol for ka. An educated scribe might try to write the word ṭuppi “clay tablet” with the special Akkadian ṭu sign, ṭu-up-pí, but ṭù-up-pí with the du signThe Hittites presumably just thought of it as du. The reading ṭù has been assigned by modern scholars to make it clearer what the intended word is, so that you can look up ṭuppi rather than duppi in a dictionary, but that doesn’t mean the Hittites actually thought of it that way!

is no less common. It’s somewhat like writing the name of the dessert as “creme brulee” versus “crème brûlée” in English: using the special accented letters is more correct in French, but most English-speakers don’t know or care, and will default to the letters they’re familiar with.

Where there are blanks in this table, the Hittites didn’t adopt any special sign for that syllable. There’s no unique sign for ṣa, for example; the Hittites just used za. And even in Akkadian, there are no unique VC or CVC signs for these consonants; iz is used equally for iz, is, and iṣ.

Akkadian prepositions

Most Akkadograms were, fundamentally, used to show off. An average scribe could write ciuc “god” with a logogram DIŊIR, or with a logogram plus phonetic complement DIŊIR-ac. But an especially educated scribe could instead write an Akkadian phonetic complement, DIŊIR-LUM, demonstrating that they know the Akkadian word for “god” is ilumOr an Akkadian and a Hittite complement for good measure, DIŊIR-LUM-ac.

.

But there were a few genuinely useful Akkadograms. Hittite had an elaborate case marking system, where nouns take different endings to show their role in the sentence. Akkadian didn’tIt had a case system, but a very limited one.

—instead, it used prepositions for that.

So these prepositions could be used with logograms, in place of the Hittite case endings! Since Hittite used spaces around words, this gave them a better way of separating logograms from the surrounding text, making them easier to read. As a general rule, phonetic signs in Hittite can never stand aloneThe main exception is the “sentence connectives”, but even they usually have other signs attached. This may be another reason to use plene spellings!

. So while DIŊIR-ni could be read as either anni or DIŊIR-ni, DIŊIR alone has to be DIŊIR.

(Of course, some scribes then added Akkadian phonetic complements to the logograms, taking away this advantage again. But then at least you get to show off your knowledge.)

Akkadian possessives

In Old Hittite, there were special possessives that could be attached to the end of a noun to indicate whose noun it was. For example, atti=mi would mean “to my father”, while atti=ti is “to your father”, atti=ci is “to his/her father”, and atti=cmi is “to their father”.

In later eras of Hittite, though, these possessives are almost never spelled outIt’s not clear if these possessives were still used at all. They’re only written out in fossilized phrases, without all the fancy inflection they had in Old Hittite. But the Akkadian possessives must have stood for something.

. In their place, the possessives from Akkadian are used instead! Just like the Hittite ones, these attach to the end of a word or phrase, and indicate who it belongs to.

In Akkadian, there are specific rules for how these should be used. =ya is used after vowels and =i after consonants, for example, =ka is only used when the “you” is male, su/sa takes the place of cu/ca after certain consonantsSpecifically, if the Akkadian word (not the Sumerian or Hittite word) ends with t, d, or ṭ, then c is replaced with s. For example, the Akkadian equivalent to É is bīt, so “his house” is É=SU, not É=CU. Proper Akkadian scribes would also change the final consonant to an s when this happened (bīs=su), which suggests that at one point, c was a fricative and s was an affricate: t /t/ + c /s/ makes ss /ts/.

, and so on.

But Hittite scribes weren’t very consistent about these things. Many of them didn’t speak Akkadian very well, so they broke these rules all the time. The only reason we use the transcriptions SÚ and SÀ instead of ZU and ZA, for example, is because of our modern knowledge of Akkadian—to most Hittite-speakers, z and s were the same sound, so they wrote it with the familiar z signs instead of the foreign s signs. So for Hittite purposes, there’s no need to bother with the details. Just recognize these signs when you come across them.

Akkadian conjunctions

There are two other especially common Akkadograms that deserve special mention.

The sign Ù, read as Ù, is used to mean “and”. This is an odd one, because it functions as a logogram! Akkadian words are normally spelled out phonetically, but “and” is specifically Ù, not U or ÚThis situation arose because in Akkadian, “and” is u, while “or” is ū. Akkadian scribes didn’t usually indicate long vowels, which was a problem here, so they created a logogram to tell them apart. The Hittites then borrowed the convenient logogram for “and”, but wrote “or” with their native word nacma to avoid that ambiguity.

. We only call it an Akkadogram because it was developed in Akkadian, not Sumerian, and we write it in italics to show that we’re talking about an Akkadian pronunciation, not a Sumerian one.

The Akkadogram Ú-UL (ŪL) is used to mean “not”ŪL is used to negate facts; the native Hittite word lē “don’t” is used to negate requests and commands.

. We know the Hittite word for this is natta, because it’s spelled out in the oldest texts, but in later eras ŪL was used exclusively. This one’s so common that it’s important to be able to recognize it.

Exercise 10

If you want to work with Hittite texts, it’s worth memorizing the prepositions and possessives, but they’re really vocabulary words rather than signs—they’re spelled out using normal Hittite CV and VC signs that you should already be used to. Do add Ù to your sign list, though.

Transcribe the following cuneiform, making note of any Sumerograms and Akkadograms.

After defeating the serpent, Inara builds a special home to keep Hopaciya in, and forbids him to ever look out the window when he’s not there. Because if he looks out the window:

nu-wa-za DAM=KA DUMU.MEC=KA a-ut-ti
“You will see your wife and your children.”

I’ve seen two different interpretations of this passage. One is that he would see his mortal wife and children, who he abandoned for the affections of a goddess, and get homesick. The other is that he would see the true form of his divine wife, Inara, and it would destroy him—compare the Greek myth of Zeus and Semele.

ma-a-an UD.20.KAM pa-it
When twenty days had passed,

a-pa-a-ca ŊIC^lu-ut-ta-an ar-ha cu-wa-i-it
He threw open the window.

Now, write the following sentence in cuneiform. Use Sumerograms and Akkadograms for the bracketed sections.

nu [his wife] [his children] aucta.
He saw his wife and his children.

When Inara comes back to the house, he begs her to let him go. The tablet is badly damaged after that, so whatever comes next is lost to history. It’s possible she lets him go back to his mortal life; it’s also possible that she kills him. Sleeping with deities seldom ends well in Bronze Age mythology, and violating the boundary between the world of humans and the world of gods is definite hubrisIn fact, part of the role of the Hittite king was to act as a bridge between these worlds.

.

Bonus: The Complaint Tablet to Ea-Nāṣir

Given the main reason people know anything about cuneiform in the modern day, I would be remiss if I didn’t include a little bit about Ea-Nāṣir. He was a copper merchant in Ur around 1750 BCE who received multiple complaints about bad customer service, which have become a bit of a meme onlineI have no idea why, but I’m certainly not complaining. Anything that makes people interested in cuneiform is a win in my book.

.

The letter is in Akkadian, not Hittite, but it doesn’t use any signs that haven’t been covered in these lessons. So if you ever want to write your own complaint tablet to the copper merchants in your life, here’s a transcription of the first third of it. Practice writing this out on clay, if possible!

a-na É.A-na-ṣi-ir
To Ea-Nāṣir
Why É instead of e? Akkadian names tend to be full sentences, usually including the name of a god; in this case, Ea-Nāṣir means “may Ea (god of magic) protect him”.Should we read this as a logogram É.A or a phonetic spelling é-a? It’s not clear. As a logogram, it would mean “house/temple of water”, which makes sense—Ea’s domain was the pure water under the earth, known as the abzu/apsû. But we know “Ea” was the Akkadian pronunciation; the Sumerians called him Enki (EN.KI), “lord of earth”. So you’ll find it transcribed both ways; the cuneiform looks the same, either way.

This could be written with a determinative, or even two (one for the person and one for the god), but at the top of a letter like this it’s unambiguously a person’s name, so Nanni didn’t bother.

qí-bí=ma
say:
Akkadian letters tend to start with a command like this. The sign qí is what you know as ki; the sign bí is what you know as ne. Since this line is so short, it gets stretched out across the entire tablet, with huge gaps between the signs.

um-ma na-an-ni=ma
Nanni (speaks) thus:

a-nu-ú-ma ta-al-li-ku
When you came,

ki-a-am ta-aq-bi-am
thus said
Aq is the same as ak or ag; remember, cuneiform distinguishes fewer consonants at the end of syllables than at the beginning. The only reason to use the reading aq here is because we know the Akkadian word for “spoke” had a Q in it.

um-ma at-ta-a-ma
you, as follows:

gu-ba-ri da-am-qú-tim
“Good ingots
Just like qí is ki, qú is ku.

a-na gi-mil-^dEN.ZU a-na-ad-di-in
I will give to Gimil-Sîn.”
Another name with a god in it: Gimil-Sîn means “may Sîn (the moon god) save him”.Sîn’s name was written EN.ZU in Sumerian, but pronounced Suen. So sometimes you’ll find the compound sign name ZUEN used to mean EN.ZU, to avoid dealing with the strange reversal. Similarly the logogram LUGAL “king”, which you saw in lesson 8, started as a combination of the signs GAL.LÚ (“big man”), from some primordial time when Sumerian put adjectives before nouns instead of after.

ta-al-li-ik=ma ta-aq-bi-a-am
You left, but what you said,
The word taqbiam “you said (to me)” was written with a plene spelling here, but a non-plene spelling in line 5. Akkadian scribes were seldom consistent about those.

ú-la te-pu-uc
you did not do.

gu-ba-ri la da-am-qú-tim
Ingots that are not good

a-na ma-ar ci-ip-ri=ya
to my messenger
Note the Akkadian preposition and possessive: to my messenger.

ta-ac-ku-un=ma um-ma at-ta-ma
you gave, saying,

cum-ma te-le-qé-a le-qé-a
“If you’ll take them, then take them,
Just like with qí and qú, qé is ke. In some times and places, the qa, qe, qi, qu signs weren’t used at all, and this was one of them.Or perhaps we should say they hadn’t been invented yet! Sumerian had no q sound, and the idea of repurposing some CVC signs to represent the Akkadian q took quite a while to catch on; otherwise, they just wrote it as the closest Sumerian sound, k.

cum-ma la te-le-qé-a at-la-ka
if you won’t take them, go away!”

ya-ti a-na ki-ma ma-an-ni-im
Me, like what kind of person

tu-ci-im-ma=ni-i=ma
do you take me as, that
The right half of this line is damaged, and the -i- sign obliterated, but we can make a good guess as to what it was.Shoutout to mostlydeadlanguages on Tumblr for helping me with this one!

ki-a-am te-me-ca-an-ni
you treat me with such contempt?

This is the first third or so of the text; after this, it continues on to list out the past grievances between Ea-Nāṣir and Nanni (in both directions!), and Nanni’s new demands. From now on he wants to inspect all the copper in person, in his own home, before paying for it, and he wants right of refusal if anything doesn’t meet his standards.

Final test

If you’re ready to test your cuneiform skills, here’s the complete myth of Illuyanka—or at least, as complete as it gets. There are still a few breaks and gaps, but this is better than most Hittite texts, since we’ve been able to piece it together from multiple copies. Most of it is presented in the original cuneiformAlbeit with obvious scribal errors corrected.

, but a couple sentences are given in transcription, for you to write out in cuneiform yourself.

The lines across the page are written on the actual tablet, separating out sections. A double line presumably indicates an especially important section break. The headers, though, are my own additions.

Signs you’ll need

SAŊŊA (SAŊŊA or SANGA) = cankunnic, “(high) priest”The only place this sign appears in this passage is actually damaged on all the surviving tablets, so it could have been a different logogram for a different kind of priest instead; different editions fill in different logograms there. But SAŊŊA is a convenient one to teach, so I’m going with that.

. This Sumerian word was actually borrowed into Akkadian as cangû and then into Hittite as cankunnic.

UM-MA (UMMA) = kiccan, “thus”. This Akkadogram is used to introduce a quotation, and appears at the top of most letters in Akkadian (“tell X that Y speaks thus:”); you’ll find it on the third line of the complaint tablet, for instance.

MUC (MUC) = determinative for snakes. This is only ever used with Illuyanka, whose name is also the Hittite word for “snake”; presumably, using d or m just didn’t feel right for the great serpent himself.Herself? Itself? Hittite nouns and pronouns don’t distinguish masculine from feminine. But Illuyanka is traditionally taken as male, based on context.

EZEN₄ (EZEN₄ or simply EZEN) = “festival”, or a determinative for festivals. We don’t know the Hittite word for this one.

ICKUR (ICKUR) = “the storm god”. This logogram has the reading IM when it means “wind”, but ICKUR (the name of the Sumerian storm god) when it’s used as a divine name. The Hittites had plenty of different storm gods, but the one in this story is Tarhont, whose name literally means “The Mighty One”. A lot of different cultures have stories about a storm god triumphing over a giant snake, which may go back to a common source.

10 (10 or U) = the number ten, which was sacred to the Akkadian storm god Adad. The Akkadians liked assigning numbers to their gods, and could use them in place of the gods’ names: 15 for Ishtar, goddess of love and war, 30 for Sîn, god of the moon, and so on. The Hittites thankfully didn’t do this very often, but they did occasionally refer to their various storm gods as d^10 as well as d^ICKUR.

LÚ.U₁₉.LU (LÚ.U₁₉.LU) = antuhcac or antuwahhac, “human”. Even though they correspond to the same Hittite word, this logogram was used instead of LÚ to emphasize the meaning “human” or “mortal” rather than just “man”. The logogram UN was also used for this word; it’s not clear why all these different logograms were used, or if there was some subtlety in the Hittite that we haven’t picked up on.

Preface

UM-MA m^ki-il-la LÚ^SAŊŊA d^10 URU^ne-ri-ik
Thus (dictates?) Killa, high priest of the Storm God of the City of NerikThat is, Tarhont.

:

ne-pí-ca-ac d^ICKUR-ac EZEN₄^pu-ru-ul-li-ya-ac ut-tar
The mythOr “story”, or “incantation”: uttar is a very versatile word.

of the Purulli-festival of the Storm God of HeavenAlso Tarhont.

.

nu ma-a-an ki-ic-ca-an ta-ra-an-zi
When theyThe priests, maybe?

speak thus,

ut-ne-wa ma-a-ú ce-ec-du
“Let the land blossom, let it thrive,

nu-wa ut-ne-e pa-ah-ca-nu-wa-an e-ec-du
“Let the land be protected,”

Nu mān māi ceczi
When it blossoms and thrives,

Nu purulliyas iyanzi.
Then they do the Purulli-festival.
Note: Remember the determinative! What sort of thing is the Purulli?

The Storm-God’s defeat

ma-a-an d^ICKUR-ac MUC^il-lu-ya-an-ka-ac-ca
When Tarhont and Illuyanka

I-NA URU^ki-ic-ki-lu-uc-ca ar-ha ti-i-e-er
stayed at the city of Kickilucca,

nu-za MUC^il-lu-ya-an-ka-ac d^ICKUR-an tar-ah-ta
Illuyanka defeated Tarhont.

Tarhontac=ta=ac=ca ciunac hōmanduc mūgait.
All the gods mourned Tarhont.
Note: Write “Tarhontac” and “ciunac” with Sumerograms and phonetic complements here. You can add an Akkadogram if you want to be extra classy.

(A break in the tablet interrupts the story here. Presumably Tarhont asked his daughter Inara, goddess of wild animals, to help him.)

nu-za d^i-na-ra-ac EZEN₄-an i-e-et
Inara made the feast.

nu hu-o-ma-an me-ek-ki ha-an-da-it
She prepared lots of everything.

GECTIN-ac DUG^pal-hi mar-nu-wa-an-da-ac DUG^pal-hi wa-al-hi-ya-ac DUG^pal-hi
The container of wine, the container of marnuwand, the container of walhi,

Nu palhac andan iyada iēt.
She put a lot in all of the containers.
Note: What determinative should “palhac” take?

Inara recruits Hopaciya

nu d^i-na-ra-ac I-NA URU^zi-ig-ga-ra-at-ta pa-it
Inara went toIt would probably be better Akkadian to say ANA “to” rather than INA “at” here, but Hittite scribes usually were not experts at Akkadian.

the city of Ziggaratta.

nu m^hu-o-pa-ci-ya-an LÚ.U₁₉.LU ú-e-mi-it
She met the mortal Hopaciya.

UM-MA d^i-na-ar m^hu-o-pa-ci-ya
InaraOne edition gives ¹Inar instead of ^dInar. This has to be either a scribal error or an editorial error.

(spoke) thus to Hopaciya:

ka-a-ca-wa ki-i-ya ki-i-ya ut-tar i-ya-mi
“Look! I’m going to do something.

nu-wa-mu-uc-ca-an zi-iq-qa har-ap-hu-ut
“YouNote how the word zikka ”you” here is written zi-iq-qa; in a later section it’s written zi-ig-ga. Since these signs were interchangeable to the Hittites, it might be more accurate to transliterate them as zi-ik-ka₄ and zi-ik-kà, but qa and ga are both easier to type and easier to recognize at a glance.

should join me!”

UM-MA m^hu-o-pa-ci-ya A-NA d^i-na-ar
Hopaciya (spoke) thus to Inara:

ma-a-wa kat-ti-ti ce-ec-mi nu-wa ú-wa-mi
“If I can sleep with you, I will come.

kar-di-ac-ta-ac i-ya-mi
“I will fulfill your wishKardiyac=tac literally means ”the (thing) of your heart”. This fossilized phrase is one of the only places the Old Hittite possessives stick around in later eras.

.”

na-ac kat-ti-ci ce-ec-ta
So she sleptThe euphemism is exactly the same in the Hittite as in the English: literally “sleep” + “with”.

with him.

The trap is laid

nu d^i-na-ra-ac m^ḫu-o-pa-ci-ya-an É-ri pé-e-hu-te-et
Inara took Hopaciya to (her) home.

N=an mūnnāit; Inaraccaz unuttat.
She concealed him, and he disguised himself as InaraOr “Inara dressed herself up”. In that case, Inara would be the one doing the luring in the next sentence.

.

N=acta Illuyankan hattecnaz carā kallicta:
He lured Illuyanka out of his holeOne tablet has hantecnaz, another has hattecnaz. We’re not sure which is the proper Hittite word for “hole”, since it doesn’t come up very often.

, (saying):

ka-a-ca-wa-za EZEN₄-an i-ya-mi
“Look! I’m making a feast!

nu-wa a-da-an-na a-ku-wa-an-na e-hu
“Come eat and drink!”

na-ac-ta MUC^il-lu-i-ya-an-ka-ac QA-DU DUMU.MEC-CU ca-ra-a ú-e-er
Illuyanka emerged, along with his sons.

nu-za e-te-er e-ku-er
They ate and drank.

na-ac-ta DUG^pal-ha hu-o-ma-an-da e-ku-er
They drank all the containers

ne-za ni-in-ke-e-er
and got thoroughly drunk.

N=e namma hattecnac kattanda nōman pānzi;
Until they couldn’t go back down into (their) hole,

Hōpaciyac=ca uit.
And Hopaciya came.

Nu Illuyankan ichimanta kalēliēt.
He tied Illuyanka up with a rope.

d^ICKUR-ac ú-it
Tarhont came,

nu-kán MUC^il-lu-ya-an-ka-an ku-en-ta
And he killed Illuyanka,

DIŊIR.MEC-ca kat-ti-ic-ci e-ec-er
And the (other) gods wereNote the VC-VC spelling here!

with him.

Hopaciya’s fate

nu-za-an d^i-na-ra-ac NA₄^pé-ru-ni ce-er É-ir ú-e-te-et
Inara built a house on a rock

I-NA KUR URU^ta-a-ru-uk-ki
in the land of TarukkaLiterally, “the land (KUR) of the city (URU) Tarukka”, like how the Hittite empire is KUR URU Hatti.

.

nu m^hu-o-pa-ci-ya-an an-da-an É-ri a-ca-ac-ta
She settled Hopaciya in the house.

N=an Inarac watarnahhickizzi:
Inara kept commanding him:

ma-a-wa gi-im-ra pa-i-mi
“When I go to the country,

zi-ig-ga-wa-ra-ac-ta ŊIC^lu-ut-ta-an-za ar-ha le-e a-ut-ti
“You must not look out through the window.
Note: Which word here means “window”? How can you tell?

ma-a-wa-ra-ac-ta ar-ha-ma a-ut-ti
“Because if you look out,

nu-wa-za DAM-KA DUMU.MEC-KA a-ut-ti
“You will see your wifeAs mentioned above, this might mean that he’d see his human family that he abandoned for Inara, or might mean that he’d see his divine wife Inara in her true form.

and children.”

ma-a-an UD.20.KAM pa-it
When twenty days passed,

a-pa-a-ca ŊIC^lu-ut-ta-an ar-ha cu-wa-i-it
He threw the window open

nu DAM-SÚ DUMU.MEC-CU a-uc-ta
And saw his wife and children.
Note: There’s an error here in the original tablet: the scribe wrote KA instead of SÚ as the third sign. What would the meaning be with KA there?

ma-a-an d^i-na-ra-ac-ca gi-im-ra-az EŊIR-pa ú-it
When Inara came back from the country,

a-pa-ac-ca ú-e-ec-ga-o-an da-a-ic
He started to cryNote that both ú and o are used for w in the same word!

,

a-ap-pa-wa-mu É-na tar-na
“Let me go back home!”

The end?

The tablet is broken at this point, so we don’t know what happened to Hopaciya after that. The next 20-some lines are badly damaged and don’t offer much clarity—something about the city of Kickilucca and the mountain-god Zaliyanu—then the next 40-some lines are missing entirely. After this, the very bottom of the tablet gives a very different version of the story.

Much less of this second version survives. In the missing section, it seems that Illuyanka defeats Tarhont and steals his heart and his eyes. The surviving part starts with Tarhont marrying “the daughter of a pauper” and begetting a mortal son, who marries the daughter of Illuyanka. He asks for the eyes and heart as a brideprice and restores them to Tarhont, who can then go out into the ocean for a rematch with Illuyanka. Illuyanka hoped that his son-in-law would protect him, but the unnamed human tells Tarhont to strike them both down, and Tarhont does.

The text ends with the line:

nu ka-a-ac a-pa-a-ac d^ICKUR-ac MUC^il-lu-ya-an-ka-ac-ca me-mi-ac
That is the story of Tarhont and Illuyanka.

Further reading

Tools

The main focus of my research at the moment is building tools to help with cuneiform studies. You’ve seen some of those in this page: the JavaScript that lets me write code like [KUR URU Ha-at-ti]{.signs} and turn it into hover-able signs, and the fonts used to render the cuneiform. You can play with it more below.

I’m working on making a page for the fonts themselves. It’ll be linked here when I do. TODO

The core technology behind all of this is known as the “kadaru encoding”, a machine-readable encoding of the shape of a cuneiform sign. The details will hopefully be published soon, but in the meantime, a preprint is available on the arXiv.

One application of this is a renderer, dubbed GAL.DUB.SAR (“chief scribe”). Type in a transliteration, and it’ll render it in various different styles, as a PNG, SVG, or PDF. That’s what I used for the Gilgamesh rendering, for example. It’s a lot more flexible than the fonts are, so you can customize the text to your taste.

Another application is a search system, that lets you look up unknown signs. Traditionally this was a tedious and time-consuming process—there’s no alphabetical order for signs (should DÀRA “ibex” come before or after BALAG “harp”?), so the only way to find the sign you want is to flip through the dictionary one page at a time scanning for itFor the computer-science-minded among you, without a total ordering, you can only do a linear search in O(N) time, not a binary search in O(log N) time.

. My goal is to let the computer do the tedious parts for you.

The traditional version of this requires typing in kadaru codes by hand, though it tries to help by showing what the sign will look like as you type. The newer (experimental) version lets you draw the strokes with your mouse. Either way, you can encode the most distinctive part of the sign, then look at every sign in Hittite that contains it. You can also use the search page directly to look up signs by name (use the “regex” box), or browse the entire dictionary this way (put in . as a regex and hit Search).

If you find these tools useful, or have suggestions for improvements, please let me know!

Hittite language and texts

If you’re interested in the Hittite language, hands-down the best introduction is Theo van den Hout’s The Elements of Hittite, which you can check out from the Internet Archive. I don’t like how it teaches cuneiform (which was part of the impetus to write this myself), but it’s a great way to learn the language, and a great reference to have on hand; it was my definitive reference when checking the examples for this page.

Olivier Lauffenburger’s “Assyrian Languages” website provides a little bit of everything, all in one convenient place: a reference for grammatical paradigms, a searchable dictionary, and (best of all for people studying cuneiform but not Hittite) a curated collection of texts. This is where I got the autograph of Gilgamesh, and is one of the resources I used for putting together the story of Illuyanka. It’s my go-to resource when I need to check something quickly.

The definitive reference on cuneiform signs is Rüster and Neu’s Hethitisches Zeichenlexikon (“Hittite Sign Dictionary”), which is unfortunately not available onlineAnd it’s under copyright, so I can’t publicly distribute my own scans of it.

. All the sign forms and readings you see on this page are based on their sketches; anyone working with Hittite cuneiform should get access to a copy, one way or another. Similar sign references exist for other languages, like Borger’s Mesopotamisches Zeichenlexikon and Labat’s Manuel d’Epigraphie Akkadienne for Akkadian, and some of those thankfully are available through the Internet Archive.

The definitive critical edition of the Illuyanka myth is Beckman’s The Anatolian Myth of Illuyanka, which I compared against Lauffenburger’s version to put together the exercises on this page. Note that Beckman’s work came before the Hethitisches Zeichenlexikon, so not all of its transliterations line up; on this page I’ve normalized them to the HZL versions.

The tangent on larger numbers is based almost entirely on Hoffner’s On Higher Numbers in Hittite, which I found fascinating enough that I had to add a section for it. It’s also why my database is the only one to have the reconstructed sign 100000 in it.

The Hethitologie Portal Mainz is the most complete online database of Hittite materials, though I find their interface extremely confusing. The CTH (Catalogue of Hittite Texts) has information on every Hittite text that’s ever been documented, and through it you can find photographs and transliterations of most of the tablets, though not autographs. All the tablet photos in these lessons come from there.

For autographs, the standard references are known as KBo (Keilschrifttexte aus Boghazköi) and KUB (Keilschrifturkunden aus Boghazköi): enormous, many-volume series with high-quality autographs of each documented tablet. Unfortunately, very few of them are online; I’ve spent many months trying to track down obscure signs in them through inter-library loan. I’m pretty sure the oldest volumes are in the public domain by now, so I should be able to post my own scans here, but I need to confirm that.

Other good resources include Kloekhorst’s Etymological Dictionary of the Hittite Inherited LexiconThough his views on the pronunciation are not exactly mainstream.

, Hoffner and Melchert’s A Grammar of the Hittite Language, Luraghi’s chapter on Hittite in Languages of the WorldTake it with a grain of salt, I found some errors in it.

, and Sturtevant’s venerableThough admittedly outdated.

Comparative Grammar of the Hittite Language. I don’t really recommend any of these for beginners, but they’re useful references once you’ve mastered the basics.

And, of course, the Ea-Nāṣir tablet. The text I used is from the Cuneiform Digital Library Initiative, which is a great resource all around for cuneiform texts. My one complaint is how low-resolution their autographs tend to be; the higher-resolution one provided on this page was found on TumblrAnd I’ve tragically lost the link. I’ll put it here when I find it again.

.

Exercise answers

Here are the solutions to all of the exercises on this page.

Answers 1

Only 1B can be checked.

(Of course, yours should be drawn on clay if possible!)

Answers 2

Nothing to be checked.

Answers 3

• ekuzi
• pati
• a-pa-ta
• a-ku

Answers 4

• kā
• apacila
• pōti
• lu-o-ri
• ce-ni
• mu-ú-li-li

Answers 5

• ha-ar-ca-an
• ic-pa-an-za
• ca-an-ah-ti
• halkic
• antuhcac
• walhzi

Answers 6

• kiccan Hōpaciyac Inari
• man=wa kattiti cec=mi nu=wa uwami
• ka-ar-di-ac-ta-ac i-ya-mi
• na-ac ka-at-ti-ci ce-ec-ta

Don’t worry about the equals signs. There’s no way to know where to put them without understanding the Hittite text.

Answers 7

• ma-ar-ri
• mar-ri
• ki-ic-ca-an
• kic-ca-an
• ag-ga-a-ta-ar
• ag-ga-a-tar

Answers 8

• pir or per “house”
• ciuc “god”
• ciwatta “days”
• utne “lands”
• wiyanac “wine”

Answers 9

• A kind of wood, or something made from wood
• A kind of vessel
• A kind of profession
• A kind of garment, or a type of cloth
• A kind of plant
• A kind of bread
• A kind of building
• A kind of bird

These are usually translated as “a kappi-vessel”, “the lakkarwan-plant”, and so on.

Answers 10

• nu=wa=za DAM=KA DUMU.MEC=KA autti
• mān UD.20.KAM pait
• apāc=a ^ŊICluttan arha cuwaīt
• nu DAM=CU DUMU.MEC=CU a-us-ta

You can also use SU instead of CU for the possessiveStrictly speaking this one should be SU, since the Akkadian word for “wife” is accat, which ends with t. But there’s no way to know this without knowing Akkadian, so using CU instead is far from unrealistic for a Hittite scribe!

, or add an Ù in between, if you like. If you look at the actual text of this section, you’ll find the scribe actually messed this up, writing DAM=KA “your wife” instead of DAM=SU “his wife”!

Answers: Bonus

This is what the actual tablet looks like, though it’s using Old Babylonian sign forms instead of Hittite ones.

Answers: Final

As before, don’t worry about the equals signs. You can’t know where to place them without a good understanding of the Hittite language. For the parts where you were asked to write in cuneiform, you might have used CVC signs where the original tablet didn’t, or vice versa, and that’s perfectly fine; what I’m transcribing here is exactly what was on the original, but there will often be other valid ways to write it.

Preface

• UMMA ^mKilla ^LÚSAŊŊA ^d10 ^URUNerik
• nepicac ^dICKUR-ac ^EZEN₄Purulliyac uttar
• nu mān kiccan taranzi
• utne=wa māu cecdu
• nu=wa utnē pahcanuwan ēcdu
• nu ma-a-an ma-a-i ce-ec-zi
• nu EZEN₄^pu-ru-ul-li-ya-ac i-ya-an-zi

The Storm-God’s defeat

• mān ^dICKUR-ac ^MUCIlluyankac=ca
• INA ^URUKickilucca arha tīērThe original tablet actually says arga instead of arha. One edition considers this a mistake for arha tīēr; another suggests that it’s a mistake for a new verb argatīēr. Either way, arga on its own isn’t a known Hittite word.
  
  
• nu=za ^MUCIlluyankac ^dICKUR-an tarahta
• d^ICKUR-ac-ta-ac-ca DIŊIR.MEC-na-ac hu-o-ma-an-du-uc mu-ú-ga-it
• nu=za ^dInarac EZEN₄-an iēt
• nu hōman mekki handait
• ŊECTIN-ac ^DUGpalhi marnuwandac ^DUGpalhi walhiyac ^DUGpalhi
• nu DUG^pa-al-ha-ac an-da-an i-ya-da i-e-et

Inara recruits Hopaciya

• nu ^dInarac INA ^URUZiggaratta pait
• nu ^mHōpaciyan LÚ.U₁₉.LU wemit
• UMMA ^dInar ^mHōpaciya
• kāca=wa kī=ya kī=ya uttar iyami
• nu=wa=mu=ccan ziqqa haraphut
• UMMA ^mHōpaciya ANA ^dInar
• mā=wa katti=ti cecmi, nu=wa uwami
• kardiac=tac iyami
• n=ac katti=ci cecta

The trap is laid

• nu ^dInarac ^mHōpaciyan É-ri pēhutet
• na-an mu-ú-un-na-a-it d-i-na-ra-ac-ca-az u-nu-ut-ta-at
• na-ac-ta MUC-il-lu-ya-an-ka-an ha-at-te-ec-na-az ca-ra-a kal-li-ic-ta
• kāca=wa=za EZEN₄-an iyami
• nu=wa adanna akuwanna ehu
• n=acta ^MUCIlluyankac QADU DUMU.MEC=CU carā wēr
• nu=za eter ekuer
• n=acta ^DUGpalha hōmanda ekuer
• n=e=za ninkēr
• ne na-am-ma ha-at-te-ec-na-ac kat-ta-an-da nu-o-ma-an pa-a-an-zi
• m^hu-o-pa-ci-ya-ac-ca ú-it
• nu MUC^il-lu-i-ya-an-ka-an ic-hi-ma-an-ta ka-le-e-li-e-etAn extra plene spelling could be used here to show that this is kaleliet rather than kalelet, but the scribe didn’t do that.
  
  
• ^dICKUR-ac uit
• nu=kan ^MUCIlluyankan kuenta
• DIŊIR.MEC-ca katticci ēcerThis is actually ēcer, not **ēcr—it’s just an irregular spelling, since the verb root is ēc- and the ending is -er.
  
  

Hopaciya’s fate

• nu=za=an ^dInarac ^NA₄peruni cer É-ir wetet
• INA KUR ^URUTārukkiWhether KUR, URU, both, or neither should be interpreted as logograms vs determinatives is an open question.
  
  
• nu ^mHōpaciyan andan É-ri acacta
• na-an d^i-na-ra-ac wa-tar-na-ah-hi-ic-ki-iz-zi
• mā=wa gimra paimi
• zigga=war=acta ^ŊICluttanza arha lē autti
• mā=war=acta arha=ma autti
• nu=wa=za DAM=KA DUMU.MEC=KA autti
• mān UD.20.KAM pait
• apāc=a ^ŊICluttan arha cuwaīt
• nu DAM=SU DUMU.MEC=CU aucta
• (With KA instead of SU, it would mean “your wife” instead of “his wife”.)
• mān ^dInarac=ca gimraz EŊIR-pa uit
• apac=ca wecgawan dāic
• āppa=wa=mu É-na tarna

The end?

• nu kāc apāc ^dICKUR-ac ^MUCIlluyankac=ca memiac

Epilogue

Acknowledgements

If you’ve gotten this far, I’d like to give a shout-out to Ryan Shosted, who got me into Hittite in the first place, and my sister Ada Stelzer, who reviewed the first version of this page and found all sorts of errors.

Bonus: A little toy

Want to display your own cuneiform the way this page does? Type a transliteration in here.

© 2024 Daniel Stelzer, released under the Creative Commons Attribution 4.0 license. Find more of my cuneiform experiments on Tumblr. Contact me at dsdraco7@gmail.com with any questions, comments, or thoughts!