Hacker News Reader

> I'm not convinced by reason #1 because words are distinct even in speech. No matter how fast or continuous your speech when you say, "The dog jumped", everyone will agree that "dog" is a distinct thing, even if you're illiterate. It seems quite logical to separate "dog" from the words before and after when you write.

Clearly you've never tried to learn a second language?

Word segmentation is a very difficult task for non-native speakers. Segmentation failure is a common error in children learning for the first time.

The word breaks are only obvious to you because you've had decades of daily practice parsing them. Look at a speech stream as a waveform or spectrogram, and they vanish: they are objectively not there.

Since several people made the same point about word segmentation, I obviously didn't express myself very well.

What I was trying to say is that, assuming you speak the language, you will recognize "dog" a separate ‘ thing’ in a sentence. If I point to a German Shepherd asking what that is, anyone who speaks English can reply "dog". It can't be shorter; it could be longer but then you're adding information ("big dog").

Even illiterate people know "dog" as a distinct ‘ thing’. Since even spoken words are distinct concepts, with a beginning and end, and having boundaries (even if the speech waveform looks continuous), it seems natural to show the boundaries (eg., by placing spaces) if you going to starting writing down the words.

Well, we know the spaces weren't added in early writing, by why not? Wikipedia's reason #1 is that adding inaudible spaces would be illogical. I'm not convinced that that was the main reason. The other reasons seem more plausible.

Well, if showing the boundaries doesn't matter for native speakers in spoken language, why would it be natural to assume that it is needed in written language?

Since we are talking about accuracy and efficiency of language here via a textual interface, I'll voice the following observation, and attempt to not be overly pedantic

The statement/question "Clearly you've never tried to learn a second language?" is an expression pattern I see frequently. A statement ended with a question mark.

Perhaps it was meant to be intentionally cheeky? Which would be funny, if I knew to interpret it that way. A /s would make this more obvious to me

Or possibly, the sentence was started based on an assumption. This was realized part way through, then a ? was tacked on at the end?

Is there another way to interpret this expression that I am not aware of?

Assumptions of others' knowledge seems to be commonplace, yet we seem to still convey ideas somewhat decently. "Rum and Coke" vs "Roman Coke". Maybe Roman's love their rum and Coke? Either way, odds are, I'll receive the correct drink.

In regards to the original article. Here are my assumptions.

I perceive a sentence as a train-of-connected-thoughts that conveys an actional expression of a person's experience, as well as their attached feelings to that experience. In that regard, I similarly view any attached grammar.

"The dog did a backflip?!" - A person receives data of an experience, first checks it's validity, realizes it falls outside their current understanding of possibility, then proceeds to double-check that they received the data correctly

> Receiver of data: "Wait, did I hear that correctly?"

"The dog did a backflip!?" - A person receives data of an experience, first accepts it (likely first-hand or from a reliable source), then realizes it falls outside their current understanding of possibility based on other reinforced models of reality, and proceeds to doubt said data)

> Receiver of data: "I didn't believe my own eyes!"

"The dog did a backflip‽" - A person doesn't have a preconceived notion for the data they are receiving, and has received it in a way that conveys that it is extraordinary - conveyed below

> Presenter of data: "The dog did a backflip!"

> Receiver of data: "Do dogs not usually do backflips?"

Feel free to correct me if I am wrong

>I'm not convinced by reason #1 because words are distinct even in speech. No matter how fast or continuous your speech when you say, "The dog jumped", everyone will agree that "dog" is a distinct thing, even if you're illiterate.

Segmenting words is actually a quite complex cognitive task that takes a while to acquire, usually considered in the framework of statistical learning[0].

Another aspect here is that the concept of a word is somewhat fuzzy, with compound words[1] emerging quite naturally, particularly in some languages. For relatively recent examples of compounds in English, consider "website" and "cellphone" (both of which are now, as a separate step, gradually becoming the default meanings of "site" and "phone").

So I don't find it too surprising that once upon a time word segmentation was even more flexible.

[0] https://en.wikipedia.org/wiki/Statistical_learning_in_langua...

[1] https://en.wikipedia.org/wiki/Compound_(linguistics)

If you listen to people speaking in a language you do not understand it can be _very_ hard to pick out the space between individual words. When learning a language your brain learns to artificially break up the almost continuous stream of sound into distinct words.

I notice this on occasion when people are speaking a language I do understand, but I miss a syllable or two when they begin speaking. The rest of their speech might as well be another language as it all runs together incomprehensibly. I cannot parse the whole sentence after missing the prompt at the beginning, and cannot even separate where the words split.

> I'm not convinced by reason #1 because words are distinct even in speech.

Not in all languages. Lookup "liaison" in French, e.g. (and a not uncommon complaint of people learning French is that they can't separate the words in spoken speech).

I'm not terrible at French and still struggle sometimes to discern the word breaks. Malapropisms in my native English can arise from that too. Learning a new language it can be v. hard to distinguish word breaks - our brain fills them in when we know a language, which is why it's hard to imagine that people might not agree that "dog" is separate.

Try saying "Hangdogears". Is it "hangdog ears" is it "hang dog ears" is it "hang dogears".

(hangdog = guilty visage; dogears = folded page corners)

Try it on someone else and see how they interpret it.

For non-native speakers, it's particularly difficult to pronounce and/or understand English composed word when they are "joined" with the same consonant (eg: newsstand, withhold, ...)

Reason #3 is not in itself a reason to forego spaces, just a reason why users would not insist on them. And no newsreader or actor today would prefer his cue cards or autocue without spaces.

I suspect a major reason not listed is inertia: people were used to reading without spaces between words, and never considered the alternative. With some help from reason #2, the increased cost.

Logical? Sure, most languages do separate words for this reason.

But languages are never completely logical. There are planty of writing systems where there are no breaks between words. Some of the ones I can think of includes Thai, Korean and Chinese.

>I'm not convinced by reason #1 because words are distinct even in speech.

Perhaps you only spend time around trained actors who enunciate their speech. Because let me tell to most people I know ramble their speech into one long string of sounds. Go listen to a recording of a language you know nothing about and tell me what the words are. Oh, and not a news cast, but everyday conversation between friends having a good time.

You speak English so you know it's "the dog jumped" and not "thud awgjumpt." But if you yell both of those quickly there's not any difference.

FORTRAN code requires no space (to save space on punchcards) ; it seems an appropriate similitude between protohistoric languages.

Amazingly, FORTRAN doesn’t even require spaces between keywords and variable names. So when the compiler sees FORI=1 it has to look still further ahead to decide if this is a for-loop (FORI=1TO9DO) or an assignment to a variable named FORI.

IIRC we actually don't pronounce most spaces unless we're being very clear, the sound is called an epiglottal stop.

I can't really do much with the numbered options there but I'd just like to point out that Thai has no spaces between words. I know six year olds who can read Thai properly so I don't know about trained performers or whatever.