An unplanned science experiment on the importance of printed spaces

Apologies if this sort of thing has been posted before. I am guessing that, as a general topic, “there are no spaces in Japanese sentences” is probably a pretty common one. I’ve personally always tried not to think too hard about it, but I have noticed that it slows down my perception when reading a new sentence. Of course, I also wondered if maybe that slowdown might be, simply, 100% the result of the foreign character set. Today, I unexpectedly answered that Q… with science? Conclusion: I believe that the foreign characters account for less than 100% of the slowdown; therefore the spaces are responsible for the remainder.

I’ve made it a habit after each review to:

  1. Read the sentence out loud myself
  2. Listen to the recording (I feel like these are usually pretty fast)
  3. Try to read it again out loud and match the speed of the recording (I rarely succeed here)

In this N4 grammar point , I got the following sentence:

お弁当を持ってきてください。

I tried task #3 from above, no exaggeration maybe 3 dozen times, and I constantly got tongue-tied in the “てきてく” region. The phrase is short enough that I basically had it memorized, but I still found that reading while reciting would make me go faster, compared to just reciting from memory (but still not fast enough).

Then I started to get mad at myself. Asking myself: if I have the sounds memorized, and they’re all sounds that my English mouth is used to making, then why the heck am I failing here? Then I got a crazy idea. I generally don’t look at Romaji, because any Romaji I read is a missed opportunity to improve my kana-reading skills. But on a whim, I asked ChatGPT (for convenience) to spit me out some Romaji:

Obentō o motte kite kudasai.

Then I tried reading that out loud. Instant success. So easy. As fast as I want. I wonder: is it because of the spaces, or because of the Roman characters? I try a new experiment:

Obentōomottekitekudasai.

I read it slower than the above version with spaces! But probably faster than the kana version.

And obviously I need to also try the below, to exercise the full range of every variable:

お弁当 を 持って きて ください。

Sure enough, I can read that faster then the version without spaces.

So to summarize, my reading speeds are as follows:

  • Romaji with spaces: easy to read at recording speed
  • Romaji without spaces: able to read at recording speed, but feels harder
  • kana with space: slightly harder than the above, but still barely able to match the speed of recording
  • kana without space: cannot consistently match the recording speed, get tongue-tied often

I started this wall of text off by saying that I’ve historically not given much thought to the lack of spaces in Japanese text. But now that I realize it is important, I probably should give it a bit more conscious thought, and perhaps work on strategies for dealing with it? One thing that I didn’t even mention above: looking at regular text (kana with no spaces) but visualizing spaces actually improved my reading speed. But I suspect that only really worked for this sentence because I had it memorized. The spaces are most problematic when encountering new text.

4 Likes

So this examples is reading outloud speed- maybe try カラオケ/歌詞. At joy joy the english songs in katakana are called Niponglish ver. I’m sure you can also find lyric videos on Youtube.
You have to read and speak at the speed of the song.
Shadowing. Netflix closed captions and read along as much as possible.

Reading comprehension is different.
I read I read the kanji first (弁当and 持)If I can understand what’s going on with just that (like in this sentence)I move on. This solves the ‘spaces’ because instead of English ‘words’ I’m breaking the sentence into ‘kanji’. Since I learned the kanji meaning stronger than the reading, I still comprehend if I don’t remember the reading

問題は、本当に真面目で熱心な教員を休職に追い込むまで精神的に疲弊させ、ひいては学校を崩壊させてしまうほど 保護者が実際に存在するということです.
I saw this sentence and 疲弊 is the only word I don’t recognize, I’m not confident of the readings for 崩壊 and保護者

OK, so I guess I’ve gotten good since this was a problem with comperhension. I’ll still go slower reading out loud because

  1. words I’m not sure of the reading and
  2. I have to process words that I read before I can say them.

My studing is 85% Jalup, 15% manga, anime and TV. I studied kanji from 2021-2022.

An interesting topic for sure.

What is “reading out loud”? That is, what is the success criteria, when should an impartial observer stop the stopwatch and compare which version of the task was completed faster?

Personally, when I try to read the romaji version with spaces above at full speed, I notice that I make the sounds first, without understanding what I’m saying. It then takes a moment to stop and think what that sentence actually corresponded to in Japanese.

Japanese version with spaces makes me want to treat 持って きて ください as 3 separate words and separate them with either a little pause or with intonation. But you have to say the whole thing on one breath and on one falling intonation. Mentally correcting for that also takes time.

On a tangent

The main character in the LN series I’m reading now (沈黙の魔女の隠しごと) actually adds awkward pauses where they should not be. In writing they express those with commas, since you cannot add spaces in Japanese.

1 Like

Reading comprehension and kanji, another interesting topic too.

I look for kanjis, but also for particles and verb conjugation endings, to isolate chunks of meaning, to understand the whole sentence.

I didn’t realize how helpful kanjis were in reading, until I did some of the JLPT practice tests in Bunpro. JLPT 5 tests are mainly in kana. I could read it, but it was difficult for me to understand it.

As you have commented -and I suppose that it happens to everyone-, sometimes I see a kanji and I know what it means, but I can’t recall its reading or even its translation in English (it’s not my native language).

Learning kanjis, I suppose that the brain makes a direct association between the “drawing” and the meaning, without passing through the reading. It helps at reading comprehension, but it’s a problem for reading aloud, for writing by typing -you can’t type kanjis- and for understanding the spoken language. And to know a language is not only being able to read it, to understand the language spoken is as important.

I’m learning kanjis in Wanikani. I wonder if the nnemonic approach reinforces this direct link between kanji and meaning or if it’s simply unavoidable, because this link is the easiest for the brain to remember. I use KameSame for extra reviews and I like it, because it also tests kanji and vocab production (writing), not only recognition. Of course, I score worse in production than in recognition.

I have an Excel sheet with the vocab that I’ve learn, for physical, offline reviews. I hid the kanji column and arranged the hiragana readings in Japanese alphabetical order. I only could guess the meaning of about 50% of the words. I only realized then how many homophones are, how many kanjis have the same on or kun reading. I know that without the context of the sentence is more difficult, but anyway…

Although Bunpro doesn’t test kanji, when a new Bunpro vocab have kanji, I must learn it, even if it is above my WK level. Because I think that it’ll be easier for me to learn the triple connection -kanji, reading and meaning- from the start, that to learn the kanji later, when I reach that WK level.

How to reinforce the kanji readings? Probably, there aren’t any specific methods, just the usual, writing, reading aloud and listening. What do you think?

Well, wanikani has reviews where it asks for readings. and includes mnumonics for reading as well. Bunpro cloze reviews practice readings as well.

This sentence reminded me about the topic of spaces or lack thereof:

大人達がそう話しているのを耳にする度、フェリクスは、
消えてなくなってしまいたくなる

消えて なく なって しまい たく なる

tl

Every time he hears adults talk about this, it makes him want to disappear.

Would this chain be more readable with spaces? Where would one even put spaces in patterns like ~てしまう, ~たい, ~なる?