I’ve asked Japanese speakers about this, but they have no answer, so I’ve turned to the sum total of human knowledge.
Japanese surnames that end in “o” (for the most part) are written as they sound in the US, for example, Ito, Kudo.
However, in Japan, the same names are spelled Itoh and Kudoh.
Any ideas why there is this discrepancy in spelling?
Ah, finally, something I know.
I believe it’s because many of those names have a double o. In Japanese, the doubling of the vowel just makes the sound longer. So Ito is pronounced “etoow”. Just doubling the o in English (Itoo) makes it look like “etwo”.
Adding the h makes more sense as to how it should sound. They also write it with a line over the vowel.
Of course, in America we don’t really care and we murder their names.
Rav
Well, I’m Japanese and my last name is Saitoh - with the “h” 
I asked my mom family about the spelling, and I was told that is is mainly just a personal preference. When we moved to the States, it was just something that my we thought was cooler 
However, my mother also told me that the “h” is also used to extend the pronunciation of the “O” at the end.
→ Saito, would end with a “sharp” “o” sound.
→ Saitoh, the end sound would be somewhat extended…Kinda like “Saitou-”
That is all I can think of!
Shpaiers
Thanks, the previous explanations I got were “I don’t know, just always been that way.”
Shpeirs: the thing, though, is that in Japanese, the syllabry doesn’t provide for an “-h” ending on any syllable. So your mom was right in describing why it’s done that way for us 'merkans who have a tendency to murder foreign names and the previous poster was right also.
In Japanese, the extension of Ito to Itoo (not pronounced ee-two!) would be Indicated by using the symbol for “I” (prounced like the “ee” in eek) followed by the symbol for “To” (pronounced like tow as in tow truck) finally followed by the symbol for “U” (pronounced like “Oooooh” as in “Oooooooh! Look at the cute baby!”.
Just a little something I learned from studying Japanese whilst stationed in the outskirts of Tokyo (NOT pronounced “Tow-kee-oh”!)
And not too long ago I was present at a discussion on the business of translations by professionals in the field, who spoke on how for some time there has been dissent (in Japan of all places) with use of the Ministry of Education’s standardized Romanji spelling (which has a specific set of phonics rules, it’s not an exact English or Latin transposition). Apparently for quite some time (pre-war) there were changing, competing, or -no- “official” standard and even after that, many people prefer their customary version; OTOH, they said, there are those who have decided that it’s cool to adopt Romanji based on the phonics of their favorite western languages – hard c’s for k’s, soft c’s for s’s, soft g’s, using Y or even ee for i, etc.
So what’s with the word ‘romaji’/‘romanji’ itself? It appears to be spelled two different ways.
And the Japanese nonreflexive tongue flap is used for either ‘l’ or ‘r’ in words from other languages, isn’t it. So how come they seem to use either letter in ramaji? Why not pick just one all the time?
And it seems like, with foreign words ending in consonants other than ‘n’, they prefer to add ‘u’. That seems like a strange vowel to pick. Does it have something to do with some kind of primacy within a syllabary?
Ray
Ray
Nano:
Heck if I know about the R/L usage, it looks pretty much arbitrarily assigned by whoever first came up with the need to transcribe word; it does look however as if the “standard” is to favor “R”.
As to the other question, under the standard rules, all syllables end in vowels; "N" being the only consonant that can end a syllable. You have a series of syllables made up of a consonant plus a sorta grunted-under-your-breath sound or stop; someone decided this is usually transcribed as "u". Thus you write as "Ketsuke" that which sounds for all the world like "Kets'ke". A good example is the Toyota model named "Camry", which happens to be close to how you pronounce its Japanese name "Kan-muri". And you get "beisaboru" for "baseball". My speculation is that "Bill Clinton" could be rendered into kana as "Bii-ru Ku-rin-toun".
These are some of those things that the translators said were causing dissent among later generations.
JR: the “N” is considered a syllable by itself.
And what about double consonants, as in Nippon, Nissan, etc. How do you get there from the syllabaries?
Ray
Also, JRD, you speak of ‘u’ as being used of a sloughed-off syllable, but doesn’t ‘i’ also end up that way, as in, say, imashita (imash’ta)?
The “doubled letters” in the Japanese Kana syllabries aren’t what those in the West would normally consider a doubled letter, such as the word letter itself. “Let-ter” being two syllables, in English; if Japan had such a word, it would be Le-tter, with the sound of the “t” being emphasized, not repeated. Ni-ppon, not Nip-pon; Ro-ppo-n-gi, not Rop-pong-ee.
The writing of the “doubled sound” is quite simple. One merely adds a circle, sort of like the degree symbol, to the syllable concerned.
Not to be nitpicky, but I can’t stop myself from correcting a couple errors in JRD’s post: it’s romaji (not romanji) and be-subo-ru (not beisaboru), where the “-” should be written over the preceding vowel to represent elongation. Also, Bill Clinton becomes Biru Kurinton (not Biiru Kurintoun). UncleBeer, though, would become Biiru Ojisan.
Monty, not quite.
First, I’ll clarify where the ‘circle’ comes into play in Japanese. There are two 48 characters syllabaries in Japanese. That’s 46 plus 2 odd ones you’re not likely to see written in modern Japanese. There are however far more syllables in the language; 89 in total (88 if ‘wo’ is simply pronounced ‘o’ as is usually the case).
Japanese uses two different symbols to change the basic syllables. These marks are usually put a little over and to the right of the character. The first symbol is called dakuten (voice mark), which look like two elongated dots. For example, the ‘k’ series, “ka ki ku ke ko”, becomes “ga gi gu ge go” (the vowels stand for their latin equivalent. Correspondingly, ‘s’ becomes ‘z’ and ‘t’ becomes ‘d’. There are exceptions, but who really wants to know? We’d be up to 48 + (3*5) = 63 syllables, but it’s actually 61 because of complicated phonetic reasons that you probably don’t want to hear.
The second symbol, the handakuten (half voice mark), which is a small circle, is used only for the ‘h’ series. If you add it to the series “ha hi hu (a.k.a. fu) he ho” you get the “pa pi pu pe po” series. If you replace the handakuten by the dakuten you get the “ba bi bu be bo” series. Now we’re up to 61 + (2*5)= 71 syllables.
The remaining 18 syllables are made by combining the syllables for ‘ki’, ‘shi’, ‘chi’, ‘hi’, ‘mi’, and ‘ni’ with the syllables for ‘ya’, ‘yi’, and ‘yo’ to obtain syllables like ‘kyo’ in Tôkyô.
Now, concerning the double consonnants. The symbol used is a smaller version of the syllable ‘tsu’ put before the syllable whose consonnant is doubled. I can WAG why this character was chosen if you’re masochistic enough to ask me. Let’s take Nissan as an example. It is written in Japanese as ni-(small ‘tsu’)-sa-n. That’s four characters, the final ‘n’ being one character.
What it actually sounds like in Japanese is another matter. Most people will agree that it makes the ‘s’ sound stronger and/or longer. Personaly, I find that it also shortens the previous vowel. It clips it, so to speak.
Momotaro, the Peach Boy
Only humans commit inhuman acts.
Oh, I wanted to add that the ‘h’ used to elongate vowels is pretty silly. It’s only useful for English speakers. It makes no sense in other languages, so I don’t like to use it. Personaly, I prefer to use a line over the vowel, which can often be rendered as a ‘hat’ on most keyboards. I think Japanese people should use a romanisation scheme to suit their own spelling, like all other languages have learnt to do.
On the other hand, some people want their name pronounced correctly. Sheesh!
Momo: Thanks. I described the wrong symbol. You’re correct in the “chisai tsu” as the doubling element.
Domo arigato, Monty and mikan, for picking up on those pitfalls of my dwindling powers of recollection.
JRD
Momo:
My last name begins with McG. It’s amazing to me the number of Americans who assume my name is pronounced “MacDonald’s!”
Yes, Momotaro, I’ve heard people ask why Americans pronounce the name of the car /nee’sahn/ when the Japanese pronounce it /ni·sahn/; so I guess the ‘ss’ does shorten/open the prior vowel.
Ray
Also, I think that Americans tend to elongate the “i” [“ee” sound] of Nissan, so that it comes out closer to “Niisan” than the “Nisan” that Japanese would say if it only had one “s”.