Writing Systems
• Pictographs (symbol looks like the thing
being represented; no major role in current
writing systems)
• Ideographs (symbols represent words or
• Syllabic writing (symbols represent syllables)
• Alphabetic writing (symbols represent
individual speech sounds)
• Morphophonemic writing (variation on
alphabetic writing – English spelling is mostly
Important Note Before We Go Through Each
Writing System
There are no writing systems in use that are
purely of any one type.
For example:
1. Chinese characters are mainly ideographs,
but with some characters specifying sound
rather than a whole word.
2. Alphabetic characters in the writing system
used for English mainly specify sound, but
ideographs are also used (&, %, $, 4, =, ? …).
What does this sign mean? The tree component
only is a pictograph – meaning conveyed directly
because it looks like the thing it represents.
Note: All four kinds or writing
are used on this sign: (1)
tree=pictograph, (2)
“SALE”=alphabetic, (3)
“4”=syllabic, (4) Exclamation
point=ideograph. (More about
this later.) But the tree is a
What does this pictograph mean?
Some particular mountains (Rockies,
Scenic view ahead?
Landscape mode (rather than portrait mode)?
Who knows?
What picture do you draw for abstract concepts
like sadden, hypothetical, creative, ambitious,
or even verbs like stall (as in “My car stalled.”)
or fidget?
What does this pictograph mean?
What do these pictures mean? How did you know?
These are ideographs – meaning is assigned by
convention; i.e., we learn to interpret them in a
particular way. In this case, the state compels us to
learn them. Also called ideograms and logograms.
Other Ideographs
Ideographs may or may not be iconic (iconic =
looks like the thing being represented). The road
signs were all iconic. The symbols above are not.
Iconic or not, meaning is assigned by convention –
we all agree that ‘%’ means “percent”, and that the
squiggly arrow means “curvy road ahead.”
Ideographs provide no clues about pronunciation;
i.e., there is nothing in the numeral “2” that tells
you to pronounce the symbol [tu].
Ideographs Used in Chinese Writing
• Any iconic elements are probably lost on most readers
• Symbols represent whole words or concepts
• No clues to pronunciation (usually)
• Both concrete (sun, river, …) & abstract words
(strength, good, peaceful) are represented
Chinese writing is not purely ideographic. Some
characters represent broad semantic categories
(e.g., person, insect, metal) and others provide
pronunciation clues.
Chinese writing is not purely ideographic. Some
characters represent broad semantic categories
(e.g., person, insect, metal) and others provide
pronunciation clues.
Writing systems derived from
Chinese are in use for Japanese
(kanji), Korean (hanja).
Downsides to ideographic writing.
1. Representation is (mostly) at the word level;
character set is really huge. The number of
symbols is in the many thousands (compare
this number to the 26 Roman letters).
2. Clues are not provided to pronunciation
(though Chinese is not purely ideographic).
Encounter a new symbol? You’re stuck. Not
true of alphabetic writing. What happened the
1st time you encountered the word
3. Place names, proper names, foreign words,
etc., can be a headache.
Downsides to ideographic writing (cont’d)
4. Specifying ideographs on a computer
keyboard requires some creativity (a one-toone key-to-symbol relationship would require
an immense keyboard.
The upside: The writing system is language
independent. Language independence means
that anyone who knows what the ideographs
mean can understand what is written – no matter
what language they speak.
This is a really big deal in China.
“Standard” Chinese: Mandarin, but a large
number of “dialects” which are more properly
viewed as separate languages. Some of the
mutually unintelligible “dialects”:
Shanghainese, Cantonese, Southern Min, Hunanese,
Northern Min, Eastern Min, Central Min, Dungan, …
several others.
Why is language independence important? A
newspaper article or book can be read equally
well by speakers of all of these languages.
This would not be true with an alphabetic writing
system, or any other system that conveyed
sound rather than meaning. This explains why
China – with many mutually unintelligible
languages – will probably never move to
alphabetic writing.
Syllabic Writing
Main idea is pretty obvious: Each symbol represents a syllable.
Japanese syllabary
(Syllabary=set of symbols
to represent the syllables
of a language)
Hiragana: One of two Japanese syllabaries.
Hiragana base characters
ん (N)
Note: The 1st row (identified with the “∅” symbol) lists
symbols used to specify syllables consisting of a vowel
by itself; e.g., the 1st syllable in the English word “Okay”.
Key difference between syllable-based
writing and logographs/ideographs:
Syllabic symbols represent sound, not
whole words.
Symbol set for a syllable-based writing
system is called a syllabary.
Syllabaries are in use for several
languages, including Japanese (two, in
fact: katakana and hiragana), Korean
(hangul), Inuit, & Cherokee.
Syllabaries are a good choice for languages with
a fairly small number of unique syllable types.
Japanese has a small number of unique syllable
types because: (1) it has a small phonemic
inventory, and (2) it has many constraints on
permissible syllable types (I’ll tell you what this
means soon).
Japanese: Somewhere around 50 unique syllable
types. (Why? Japanese has just 5 vowels, a little
over a dozen consonants, no consonant clusters;
syllable types consist mainly of V and CV.)
English: Many thousands of syllable types (~3
times as many vowels as Japanese, ~twice the
number of consonants, lots & lots & lots of
consonant clusters. For example, a separate
syllabic symbol would be needed just to
represent the word sphinx. That symbol would
have no other use. Same thing with the word
Would syllabic writing be a good choice for
English? Not impossible, but it would be
Alphabetic Writing
Sound is represented rather than whole words,
but sound is represented at the phoneme level,
not the syllable level.
“Ideal” alphabetic system: 1 letter = 1 sound. No
system in use meets this ideal, but some are fairly
English isn’t one of the close ones. (Big surprise,
Roman Alphabet
English (along with many other languages) is written
using the Roman alphabet.
Lineage (i.e., where the alphabet come from):
(1) Hieroglyphics -> West Semitic Syllabary
(Phoenicians – hence the word phonetics)
Hieroglyphics was a mix of ideographs &
syllabic symbols; Phoenicians ditched the
ideographs and kept a 22-symbol syllabary.
(2) West Semitic Syllabary -> Greek Alphabet
Greeks added some vowel symbols and turned
the syllabary into an alphabet.
Origin of Roman Alphabet (cont’d)
(3) Greek Alphabet->Roman Alphabet
Romans added a few symbols and redefined
others to suit the phonetic inventory of Latin.
Quick Summary
1. Hieroglyphics (mix of ideographs & syllabic symbols)
2. W. Semitic Syllabary (keeping syllabic symbols only)
3. Greek alphabet (switch from representing sound at the
syllabic level to the phoneme level)
4. Roman alphabet (a few changes to better fit the
phonetic inventory of Latin)
From Simpson Bible Stories
Pharaoh (Principal Skinner) dictating a
letter in hieroglyphics: “… giant eye, dead
fish, cat head, cat head, feather, picture of
a guy doing this …”.
English Spelling Stinks Big Time
English is not the only spelling system that
is problematic due to inconsistent letter-tosound relationships. Many spelling systems
suffer from this problem. But English gets
the prize as the worst, at least here on earth.
Check out this link:
Why English Orthography Stinks
1. Roman alphabet is poorly suited to English for
a very simple reason: The sound inventories of
the two languages are very different, especially
the vowel inventories.
Latin: 5 monophthongs, no diphthongs
English: ~12 monophthongs, 3 diphthongs = 15
How do you represent 15 vowels with 5 symbols?
*** NOTE: The poor fit of the Roman alphabet to
the phonetic inventory of English is the 1st item
on this list, but it is far from the most important
problem. (More to come.) ***
Problem with the Roman alphabet could be dealt
with by consistently using combinations of letters
to represent sounds. A lot of this is done:
met–meat, bit-beat, cot–coat, ...
But: it isn’t done w/ much consistency. Letter-tosound associations are all over the place.
/i/: ski, flee, meat, people, physiology, fetus, ...
Same problem with consonants:
/f/: foam, phone, rough, etc.
Many other examples. G.B. Shaw:
“fish” = ghoti (gh as in cough, o as in women, ti as in
2. There are sounds in some words that could
easily be represented with the Roman alphabet
but, for one reason or other, they are not.
beauty: What speech sound follows the [b]?
user: What speech sound begins this word?
union: What speech sound appears twice in this
word but is not represented at all in the
The “stealth” sound is [j] in each of these words
([bjuɾi], [juzɚ], [junjən]), but English spelling
provides many other examples of sounds that
could be represented with the Roman alphabet,
but are not.
Moral: The big problem isn’t really the Roman
alphabet. The alphabet wasn’t designed for
English, but the liabilities could be dealt with.
They aren’t. How’s come?
The Real Culprits
1. Spelling standardization is surprisingly recent,
and it was never done in a systematic way. As
recently as the 19th C even educated writers
spelled (NOT spelt) however they felt like.
When spelling was finally standardized, it was
done in a haphazard way. So, spelling was
standardized but not systematized.
Systemizing spelling is what was really needed,
but this wasn’t done.
Quick note:
On exams I sometimes ask students to list and
explain the most important problems with
English spelling.
Q: What is wrong with the answer “spelling
standardization” (which I get quite often)?
A: It does not answer the question. Suppose I
had told you that a major problem with
English spelling was spelling standardization,
then moved on to the next one. Would this
have meant anything to you?
Lesson: You do not need to write a book, but you
should aim to capture the key idea.
The Real Culprits (cont’d)
2. Historical sound change: Pronunciation
changes over time; spelling usually stays put.
knight, knife – the ‘k’ ([k]) used to be pronounced
gnaw, gnat – same deal with the ‘g’ ([g])
cough, rough – words used to end with a speech
sound that is no longer in the phonetic
inventory of English – the consonant that is at
the end of the name “Bach”, as it is spoken by
3. ”Foreign born” words. Most English words
started out life as part of the vocabulary of
some other language. The habit is usually to
retain the original spelling.
pizza, colonel, cello, junta, chassis, psalm,
repertoire, liaison, sauna, coup, algorithm,
decipher, coffee …
Result: Mishmash of different and wildly
inconsistent spelling conventions.
This is almost certainly the most important of
the “culprits”.
4. Others – See MacKay (Chapter 3), but the three
points discussed above are the main ones. These
are the ones I’d like you to know (& be able to explain).
For the Etymologically Curious
French (who borrowed it from the Italian
colonnello, meaning ‘little column’)
OK, so English spelling has problems
everywhere you look.
These problems explain why:
 It took most of us forever to learn to spell.
 Most of us still have problems from time to
time; more than occasionally for some of us.
 It’s unusual to find people who consider
themselves to be good spellers. It’s about as
common as finding someone who says, “Yes,
I find that algebra comes very easily to me.”
BUT: Is English spelling really as bad as all that?
Not really.
The last topic we’re going to discuss addresses
exactly this question: English spelling is a mess,
no doubt about. But is it really quite as messed
up as it seems? The answer is NO.
The fact that English spelling is not phonetic is
NOT a design flaw. English spelling is
intentionally not based on a phonetic design
principle; i.e., it is purposely not trying to
represent surface phonetic details.
Instead, it is based on something called a
morphophonemic principle. This is the most
important concept in the entire section on writing.
A quick preview of where we’re heading with this.
In spite of all of its many and very real flaws,
there is an elegance to our spelling system (and all
other alphabetic spelling systems).
It’s called the morphophonemic principle. Details
very soon, but here’s the basic idea:
The spelling system is purposely designed NOT
to represent surface phonetic details.
Instead, it’s designed to represent more abstract:
(1) morphological regularities, and (2) phonemic
(not phonetic) regularities.
IMPORTANT NOTE (last thing)
The morphophonemic principle is not a hard idea,
and pretty much everybody gets it.
So, here goes ...
Two starting points:
1. English spelling could not be made strictly
phonetic even if we wanted it to be.
2. We would not want it to be strictly phonetic.
1. It Could not make phonetic even if we wanted it
a. No solution to dialect variation. Whose dialect
should we use? British? Australian? Irish?
Scottish? S. African? American? Suppose we
chose American. Which dialect? Brooklyn?
South Philly? Rural Alabama? Looosiana? We
wouldn’t want spelling to vary across these
communities – i.e., you’d want a Brit to be
able to read a book written by an Aussie. This
problem is not solvable.
b. No easy solution to changes in
pronunciation patterns over time. Speech
patterns change over time – and much more
quickly than most people imagine.
How often do you want to reprint your
So, it would be somewhere between difficult and
impossible to make English spelling phonetic.
Q: Would we want spelling to be phonetic?
A: No. No doubt about it.
Why? Because morphophonemic spelling is a
much better idea than strictly phonetic spelling.
Major Point #1: English spelling represents underlying
morphological regularities, not superficial phonetic
facts. Simpler than that sentence makes it sound.
approximate, approximately
Note the highlighted letter “a” (NOT /a/ or [a]). Is it
pronounced the same way in the two words? [No]
Would you want the spelling to: (1) change to reflect
the difference in pronunciation? or (2) stay the same
to reflect the fact that these are two versions of the
same word, related to one another by a morphological
I vote for #2.
This is not a minor phenomenon in English spelling.
It’s all over the place.
nation – nationality
medicine - medicinal
magic – magician
social – society
discuss – discussion
televise – television
(different sounds [e,æ], same letter)
(different sounds [ɛ,ə], same letter)
(different sounds [k,ʃ], same letter)
(different sounds [ʃ,s], same letter)
(different sounds [s,ʃ], same letter)
(different sounds [z,ʒ], same letter)
In all cases the spelling reflects not phonetic details
but the deeper and more important fact that the word
on the right is derived from the one on the left by
morphological rules. This is the ‘morpho’ part of
English spelling is morphophonemic, not phonetic.
This is a good thing, not a bad thing.
Major Point # 2: English spelling represents underlying
phonemic regularities, not surface phonetic facts.
walk – walks
cat – cats
dog – dogs
pan – pans
Phonological rule: If the word ends in a voiceless
consonant, add [s]; otherwise, add [z]. But note that the
spelling does not reflect this – always get ”s” in spelling.
On an abstract, phonological level, it’s an /s/ in both cases:
a sound-pattern rule handles the phonetic detail of
whether this “abstract” /s/ is realized as [s] or [z].
Similar to the morphology thing: The spelling system
reflects underlying linguistic (phonemic/phonological)
facts, not superficial phonetic facts. This
is good.
Point #2 (cont’d): English spelling represents
underlying phonemic regularities, not surface
phonetic facts.
walk – walks
cat – cats
dog – dogs
pan – pans
The example in the previous slide (shown above)
illustrates the phonemic part of the morphophonemic principle.
Point #2 (cont’d):
This is very smart. Why? There is no need to
use the spelling “dogz,” even though it is
phonetically more accurate. The ‘z’ is not
needed because the reader already knows that
the word is pronounced [d
ɔgz]. How? Even very
young kids have already learned this rule.
So, the information that the spelling system
fails to provide to the reader (whether the plural
marker is pronounced [s] or [z]) is something that
they already know. The unvarying “s” in
spelling quickly tells the reader “this word is
plural,” which is all they need to know.
Example 2:
backs – bags ([bæks] - [bægz])
Spelling rules could do one of two very different things:
(1) Use an “s” for “backs” and a “z” for “bags” to reflect the
difference in phonetic details (phonetic spelling), or
(2) Use the same letter in both cases so that the reader can
easily see that all that is happening is that the base
morpheme (back or bag) is being made plural
(morphophonemic spelling).
Which kind of system is actually in use in English?
Example 3: jogged - walked
Note that the spelling shows a “d” in both cases;
however, in “walked” the final consonant is actually a
[t], while in “jogged” it is a [d].
To work out on your own:
1.In what way is example 3 (previous slide –
jogged vs. walked) analogous to example 2?
2.In what way is this another example of
morphophonemic rather than phonetic
Types Writing Systems
(Note: No system is purely any of the types listed)
Symbols Represent
Words or Concepts
Meaning is (meaning by convention)
Supposed to
be Obvious
Chinese, Japanese (kanji),
Dead Some symbols used in
End English (! @ # $ % 1 2 3 …)
Symbols Represent
Japanese (kana)
Korean, Inuit,
English, many
SUMMARY (cont’d)
Which kind of alphabetic system is actually
in use in English? Phonetic or
The International Phonetic Alphabet (IPA)
In one important sense the IPA does not
belong in this section – it is not used as the
writing system for any language. No books
written in IPA, no newspapers, people don’t
write letters to one another in IPA.
IPA was developed explicitly for the study
and description of the sound patterns of the
world’s languages.
The IPA has two important features:
1. This alphabet really is phonetic (and sometimes
phonemic) – one-to-one correspondence
between speech sounds and symbols: The
symbol /i/ always specifies the vowel in “beet”;
every time the vowel in “beet” occurs the
symbol /i/ is used to represent it.
2. Designed to represent the sounds of all of the
world’s languages.
Does the IPA represent sound at the phonetic level
or the phonemic level?
Example: How would we transcribe the words “cat”
and “scat” or “pot” and “spot” or “bat” and “ban”?
Answer: The IPA allows you to transcribe either at
the phonetic level (e.g., representing phonetic
details such as aspiration or vowel nasalization)
or at the phonemic level (i.e., ignoring phonetic
details and specifying only the phonemic
Phonetic: [khæt]-[skæt], [phɑt]-[spɑt], [bæt]-[mæ̃n]
Phonemic: /kæt/ - /skæt/, /pɑt/-/spɑt/, /bæt/-/mæn/
[khæt]-[skæt], [phɑt]-[spɑt], [bæt]-[mæ̃n]
Phonemic: /kæt/ - /skæt/, /pɑt/-/spɑt/, /bæt/-/mæn/
How are these different?
• Phonetic transcription represents fine phonetic details;
e.g., the allophonic differences between [kh] and [k], [ph]
and [p], [æ] and [æ̃]. How many details? It’s up to you.
• Phonemic transcription ignores the fine phonetic
details, representing only the broad phonemic
categories /k/, /p/, and /æ/.
• Phonemic symbols are enclosed in slashes (e.g., /k/, /p/,
• Phonetic symbols are enclosed in square brackets (e.g.,
[kh], [k], [ph], [p], [æ], [æ̃]).
Which system are we using in this course?
Phonetic? Phonemic?
We’re using a middle-of-the-road system called
broad phonetic transcription. All this means is
that we’ll mark some allophonic details but not
Which ones? Unless I specifically tell you
otherwise, do it the way we’ve been doing it in
class and in lab.
This is not new. You now have a name for it.
Q: In this course, which allophonic details do you
need to mark?
A: The ones we’ve been using in class. There’s
nothing that needs to be memorized.
(flap, an allophone of /t/)
(glottal stop, another allophone of /t/)
(schwa, an allophone of all English vowels,
controlled by stress – if unstressed, you get schwa;
more later)
For later: syllabic consonants; e.g., the /n/ in:
button [bʌʔn̩] cotton [kɑʔn̩] (Note: The diacritic beneath the ‘n’
is supposed to be a vertical line centered beneath the symbol. It does
not print correctly. No idea why.)

writing - Western Michigan University