41

Today my 9-year-old nephew told me that he can spell any integer in English using only 9 letters. This is how he's doing it:

    ...
-3: MINUS ONE MINUS ONE MINUS ONE
-2: MINUS ONE MINUS ONE
-1: MINUS ONE
 0: ONE MINUS ONE
 1: ONE
 2: ONE PLUS ONE
 3: ONE PLUS ONE PLUS ONE
    ...

The letters he's using are E, I, L, M, N, O, P, S and U.

Can we do better and spell any integer with less than nine letters? I think so!

Just please be sure to briefly explain any maths you are using in your method, so that my young nephew (and his uncle) can understand.

Rosie F
  • 8,621
  • 2
  • 18
  • 50
GOTO 0
  • 13,549
  • 6
  • 47
  • 92
  • Well, the obvious (probably invalid because of missing brackets) answer would be to just use MINUS MINUS instead of PLUS, resulting in 7 characters. –  Sep 10 '16 at 10:37
  • It'd be interesting if other languages could be allowed. – Shuri2060 Sep 10 '16 at 14:24
  • 18
    I,V,X,L,C,D,M - no negatives though. – Mazura Sep 11 '16 at 01:02
  • 4
    Of course, the question taken literally as asked can be done with eight letters - to spell the literal phrase (A N Y I T E G R). Granted, that's not what it meant to ask (and the answers are good)...but my interpretation isn't specifically forbidden in the question, either :) – Megha Sep 11 '16 at 02:37
  • 7
    @Mazura If you're going to stretch the definition of "spell" to include non-English spellings, then why not just use I? 10 is IIIIIIIIII for example. Then apply an encoding of zero and negative numbers so I is 0, II is 1, III is -1, IIII is 2, IIIII is -2, IIIIII is 3, ... – user253751 Sep 11 '16 at 07:42
  • 3
    I is 1, II is 2. rotating 90° I is minus. Just one letter. – Takahiro Waki Sep 11 '16 at 18:06
  • 5
    This seems really ambiguous, and the variety of answers seem to support this. What exactly does it mean to "spell any integer in English"? – Jason C Sep 11 '16 at 23:30
  • 8
    I can do with zero letters, by simply using digits and minus sign instead, in the usual way. ;-) – celtschk Sep 11 '16 at 23:48
  • I would suggest an edit that changes "spell" to "write" – Kevin Sep 12 '16 at 14:39
  • "MINUS ONE" is not a number, because "minus" indicates subtraction. He could say "ONE MINUS ONE MINUS ONE" etc. and get the result of "negative one", however. – Monty Harder Sep 12 '16 at 18:39
  • If you were to consider each segment of a seven-segment display to be a single line (represented by the letter I), you could make any number, positive or negative, using just one letter. Throw in a dot for decimals and you're still only up to two "things". – Ian MacDonald Sep 12 '16 at 19:04
  • @MontyHarder MINUS ONE indicates -1. The vast majority of Mathematicians and users of the English language would agree on that. – Shuri2060 Sep 12 '16 at 19:29
  • I think you need to clarify what it means to "spell" an integer. I could just use the English words "A" and "I", and encode any number in binary. Do you mean the resulting integer encoding is some sort of instruction to follow to generate the integer? – GManNickG Sep 12 '16 at 20:39
  • @GManNickG Well, I think if I encoded a number that way, I would also need additional instructions to explain how my encoding works, and I'd probably need more than two unique characters to spell those instructions. By spelling I just mean "writing with letters". – GOTO 0 Sep 12 '16 at 20:51
  • 3
    I'm closing as "too broad" since the validity of many of the answers seems to be opinion-based. There's no way to find an objectively correct best answer. – Deusovi Sep 13 '16 at 02:17
  • @Deusovi why not use the "correctness of the answers is opinion-based" close reason? – ffao Sep 13 '16 at 19:20
  • How would you do this in French? – JMP Mar 23 '19 at 11:04

16 Answers16

43
  ...
-3: MINUS ONE MINUS ONE MINUS ONE
-2: MINUS ONE MINUS ONE
-1: MINUS ONE
 0: ONE MINUS ONE
 1: ONE 
 2: ONE MINUS MINUS ONE
 3: ONE MINUS MINUS ONE MINUS MINUS ONE
    ...

two letters fewer since we do not use PLUS any more.

Oray
  • 30,307
  • 6
  • 61
  • 215
  • +1 Was about to post this when I read the question, but you beat me to it a few hours ago. – Kevin Cruijssen Sep 10 '16 at 13:15
  • Is this considered unambiguous? For example, the expression 1 - - 1 would generally not be considered valid (but 1 - (-1) would). – svick Sep 11 '16 at 10:40
  • @svick I think there's only one interpretation possible of ONE MINUS MINUS ONE: 1 - (-1) – Shuri2060 Sep 11 '16 at 10:41
  • 1
    You do have to be careful that ONE MINUS MINUS ONE MINUS MINUS ONE isn't misheard as $1 - (-1 - -1)$, but I think that's doable. – Ben Millwood Sep 11 '16 at 11:51
  • 3
    I think $a-b-c$ is well defined as $(a-b)-c$ (grouped left to right). – Kos Sep 12 '16 at 09:40
  • @Kos Not generally considered true; actually, by the standards of most programming languages, it is true; by the standard of mathematics, it is undefined. (In general, $a \operatorname{op}_1 b \operatorname{op}_2 c$ is well defined only if $\operatorname{op}_1$, $\operatorname{op}_2$ have defined precedence or if the operations are mutually associative within the given context, i.e. if it doesn't matter which operation you perform first.) – yo' Sep 12 '16 at 10:47
  • 1
    @yo' No, I think Kos is right. Writing $a-b-c$ is acceptable and is interpreted as $(a-b)-c$ even if this may not necessarily be true for operations other than subtraction. – Shuri2060 Sep 12 '16 at 12:17
  • @QuestionAsker "right" for the use in an answer to a puzzle? Quite sure. Right for general use without explanation, remark or even a definition? Not so quite. If I was to review a math paper and it used this, I would ask for clarification. – yo' Sep 12 '16 at 12:22
  • @yo' Surely we assume BODMAS (or BIDMAS, etc) whenever we write something like that. It says we apply the operations in the order specified and from left to right. Since we have two of the same operation here, we apply it from left to right and get $(a-b)-c$? – Shuri2060 Sep 12 '16 at 13:14
  • @QuestionAsker Sorry, I'm a mathematician; the "only" assumption I make is that the axioms are consistent. All the rest is a matter of definitions and proofs. – yo' Sep 12 '16 at 13:22
  • @yo' So am I... although I'm happy to follow widely accepted definitions. :) I see your point, however, and it does emphasize the necessity for rigorous definitions and notation to eliminate all ambiguity. – Shuri2060 Sep 12 '16 at 13:25
  • I know this isn't english.stackexchange.com, but it the phrase really is two letters fewer – Eris Sep 12 '16 at 15:55
  • The words "minus" and "negative" are not the same thing. There is no number "minus one"; only "negative one". "Minus" refers to the process of subtracting a number from another number. – Monty Harder Sep 12 '16 at 18:40
  • @MontyHarder Yeah, but it's English, a strange and fuzzy language that's generally disconnected from technical correctness, and if you went up to somebody on the street and said "minus one" you'd put a "-1" in their head, which is the correct result. Do it enough times and we'd eventually just update our dictionaries, ha. At least this answer is one that doesn't need further explanation, most of the other answers use these strange encodings that you couldn't just say to somebody without giving them the full background first. – Jason C Sep 13 '16 at 13:01
25

4 letters: E N O T

Explanation:

My idea is to use binary and marcoresk's idea of using NOT to represent the - sign, although I have to admit it might be slightly 'cheating' as I have to define things and work in a different base.

To represent a number you read out its binary representation using ONE and NONE to represent $1$s and $0$s.

Eg. $5$ would be: ONE NONE ONE

If the number is negative, add a NOT to the beginning.

Eg. $-6$ would be: NOT ONE ONE NONE

Extension:

You can represent any decimal which is terminating in binary with an extra letter D for a total of 5 letters which allows you to say DOT for the decimal point.

Eg. $-5.5$ would be: NOT ONE NONE ONE DOT ONE

Other possibilities:

I understand that using NOT probably isn't the best idea to indicate -, so as I've mentioned in the comments, perhaps NEG can be used instead as it's an abbreviation for negative (I wasn't sure of this until Jason C pointed it out in his answer). This will still result in 4 letters being used (although 6 for the extension).

Shuri2060
  • 3,068
  • 16
  • 40
  • You could define a negative bit and get rid of the T – Carl Sep 11 '16 at 04:22
  • 6
    I think NOT x should be interpreted as -x-1 rather than -x. – R.. GitHub STOP HELPING ICE Sep 11 '16 at 04:50
  • @R.. Positive: ~x is -(x+1), not -x. – EKons Sep 11 '16 at 09:20
  • I would think NOT is 0 or FALSE, not - – Raystafarian Sep 11 '16 at 12:28
  • 1
    Well — I was only putting a suggestion for defining NOT which sort of makes sense. I wasn't thinking about the actual Boolean definition. Would NEG work for negative instead? – Shuri2060 Sep 11 '16 at 13:17
  • Of course, I could just take Jonathan Allan's idea of having the first bit represent the sign like computer data which would use only 3 letters. – Shuri2060 Sep 11 '16 at 15:01
  • I'm pretty sure the intent of the question was that the meaning would be clear by some ordinary interpretation of the natural-language words used, not whether you could somehow encode meaning in the letters where it would require knowledge of your scheme to interpret. Otherwise the answer is obviously one: express an enumeration of the integers in base-one. – R.. GitHub STOP HELPING ICE Sep 11 '16 at 15:42
  • There is precedent for the word "not", as an arithmetic operator, to have "not x" mean "-x-1". I can't think of any other meaning you could reasonably expect it to have if applied to integers and not booleans; certainly a definition that yields "not zero = zero" is not reasonable. – R.. GitHub STOP HELPING ICE Sep 11 '16 at 15:44
  • I don't understand how your definitions make sense, though. Could you provide a link? If I plug 0 or 1 (which usually represent TRUE/FALSE in Boolean) I don't get the other? – Shuri2060 Sep 11 '16 at 17:04
  • I could also use Ali Humayun's idea of using OFF as subtraction. So for any negative numbers I would say: 'the magnitude' OFF NONE. This would keep the 4 letter total. – Shuri2060 Sep 11 '16 at 21:22
  • Dot one means ".5"? – Canadian Luke Sep 12 '16 at 02:18
  • If we're using binary and iffy definitions, why not just use "on" for 1 and "no" for 0, then you could do the whole thing in 2 letters. Keeping in mind he said "Integer" so we don't need a decimal point. – Jae Carr Sep 12 '16 at 03:58
  • @CanadianLuke Yes. The $n$th place after the decimal represents $2^{-n}$. It was just an extension which I thought I might add because it used only one more letter. – Shuri2060 Sep 12 '16 at 10:01
  • People read "101" as "one o one". Why not replace "none" by "o"? Same amount of distinct letters, just shorter ;) – Olivier Grégoire Sep 12 '16 at 11:48
  • @R.. "not x" = "-x-1" is two's complement binary numbers. In one's complement "not x" = "-x". – Johnbot Sep 12 '16 at 14:13
  • @Johnbot: The naming is deceiving because it makes it sound like "twos complement" is something special. Rather it's just "math" as opposed to "some engineer's wacky idea". :-) This is reflected in the fact that the bitwise operators in C are not really well-defined for signed operands on (purely theoretical) non-twos-complement implementations (complex topic and probably off-topic here). – R.. GitHub STOP HELPING ICE Sep 12 '16 at 14:44
15

If we use LESS to indicate subtraction, then we have

-1 = ONE LESS ONE LESS ONE

0 = ONE LESS ONE

1 = ONE

2 = ONE PLUS ONE

etc., using only seven letters ELNOPSU.

Rosie F
  • 8,621
  • 2
  • 18
  • 50
  • What about using punctuation? Say "ONE LESS: ONE LESS ONE LESS ONE" as 2 (1 less -1). That gets you to 5 letters. Or 6 if you count the colon, which you really shouldn't. – Hugh Meyers Sep 10 '16 at 18:52
  • @Hugh Punctuation strikes me as cheating. – Rosie F Sep 10 '16 at 20:48
  • I don't think you need punctuation; it's clear enough that "less" is being used as an operation. – Samthere Sep 12 '16 at 12:31
  • @Samthere "Less" is indeed being used as an operation -- I agree with you there. But Hugh's intended solution needs to distinguish between 1-1-1-1 = -2 and 1-(1-1-1) = 2. The 4 operands and 3 operations are the same in the two cases. But they are grouped in different ways in order to get different results. And it's the proposed means to do that which struck me as cheating. – Rosie F Sep 12 '16 at 12:36
  • @RosieF Oh, I see what you mean; Hugh's suggestion is for using your subtraction operator to achieve positive numbers instead of having an addition operator. I agree that it's cheating, or rather that to do it in the spirit of the question would require you to write "OPEN BRACKET" "CLOSE BRACKET" or something similar. – Samthere Sep 12 '16 at 13:04
13

If any integer can be spelled with this structure

one some operation one some operation one

the trick could be simply this: reduce the numbers of letters needed for operations (or use O, N, E as much as you can!).

M,I,N,U,S introduces 4 new letters (N does not count) P,L,U,S introduces 2 other letters (P and L)

A simple (but symbolic) solution is to represent a operation with a single letter... but this is not a way to "say it in English".

So I propose this structure: PLUS becomes AND (only two more letters, A and D). Every integer number N bigger than 1 can be said as

N =  ONE AND ONE AND ONE AND ONE...

B) Define the number MINUS ONE as "NOT ONE" (let's say this will became our convention, this is the weakest part!) in order to introduce only one more letter (T)

Now zero becomes

ZERO =  ONE AND (NOT ONE)

C) Since every negative numbers is equal to its absolute value multiplied by -1 e.g. -59 = 59 * (-1) I suggest to use a new word to introduce the operation of multiplication, as in class we often say A multiplied by B as "A dot B" (A dot is one possible symbol to show multiplication in advanced math, instead of x or * or others)

Note that now DOT does not introduce any new letter. So -3 could be

(ONE AND ONE AND ONE) DOT (NOT ONE)

Notice we have used only 6 letters instead of 9 (and mantaining a speakable structure) O, N , E, D, T, A

Now the last problem: how to "say" parenthesis? With the rithm of speech! ONE AND ONE AND ONE AND ONE [after this sequence do a little pause] DOT [say DOT as an important word, then a little pause] NOT ONE [say NOT ONE fast, as if would be a single word]

Thank you for this pretty puzzle and sorry for my poor english!

marcoresk
  • 1,057
  • 8
  • 10
  • How about defining NOT as an operator which multiplies by -1? You wouldn't need DOT in C), although it doesn't make a difference to the letter count. – Shuri2060 Sep 10 '16 at 13:04
  • 1
    Welcome to Puzzling! – IAmInPLS Sep 10 '16 at 13:08
  • @QuestionAsker your doubt is right, thank you for asking. I made this choice basically because i wanted to force my answer to be "the more speakable" as possible (it was the request after all). And introducing the DOT I didn't add any letter, so I preferred a longer but more understandable (I hope!) solution – marcoresk Sep 11 '16 at 09:30
  • Why not have ZERO = NONE? – SQB Sep 12 '16 at 10:55
  • @SQB I mainly use the zero to exemplify the NOT operator, needed for negative integers, and simply because the using of NONE does not reduce the 6 letters ;) – marcoresk Sep 12 '16 at 12:07
9

Since

"O" is in the Collins English dictionary as "used to mean zero"
(pronounced the same as "owe", as in "I owe MackTuesday a coconut")

...and, as pointed out by @MackTuesday, "a" means one
(as in "I owe MackTuesday a coconut")

We can

Use just two letters: O, and a
(I previously used three with One rather than a, which may be less confusing)

If we use base 2, or binary with a signed magnitude representation by using our first bit as an indication of "is this number negative" and the time we stop speaking to represent our most significant bit for example:
$-32 =$ a O O O O O a $=- (0\times 2^0 + 0\times 2^1 + 0\times 2^2 + 0\times 2^3 + 0\times 2^4 + 1\times 2^5)$

$-7 =$ a a a a
$=- (1\times 2^0 + 1\times 2^1 + 1\times 2^2)$

$-4 =$ a O O a
$=- (0\times 2^0 + 0\times 2^1 + 1\times 2^2)$

$0 =$ O O (or a O)
$=+ (0\times 2^0)$ (or $=- (0\times 2^0)$)

$17 =$ O a O O O a
$=+ (0\times 2^0 + 0\times 2^1 + 0\times 2^2 + 0\times 2^3 + 1\times 2^4)$

Note that

$2^n$ just means $2$ multiplied by itself $n$ times, so $2^3$ is $2\times 2 \times 2=8$
...and that the identity element of the multiplication group is $1$ so $2^0$ is $2$ multiplied by itself $0$ times, which must be $1$.

Jonathan Allan
  • 21,150
  • 2
  • 58
  • 109
9

6 Letters: T I M E S U

-1: i times i
0: sum i i times i times i
1: i times i times i times i
2: sum i times i times i times i i times i times i times i
3: sum i times i times i times i i times i times i times i i times i times i times i
-2: i times i times sum i times i times i times i i times i times i times i

i being the mathematical constant for the square root of -1. i squared ("i times i") is -1 and i to the fourth power ("i times i times i times i") is 1.

If the pause between the items in the set it not clear enough, you could also use the notation:
3: sum set item i times i times i times i item i times i times i times i item i times i times i times i

Kys
  • 191
  • 3
8

Another way of using only seven letters (A, D, E, L, N, O, S):

    ...
-3: ONE LESS ONE LESS ONE LESS ONE LESS ONE
-2: ONE LESS ONE LESS ONE LESS ONE
-1: ONE LESS ONE LESS ONE
 0: ONE LESS ONE
 1: ONE
 2: ONE AND ONE
 3: ONE AND ONE AND ONE
 4: ONE AND ONE AND ONE AND ONE
    ...
Rand al'Thor
  • 116,845
  • 28
  • 322
  • 627
8

ONE ON ONE = 2

ONE OFF ONE = 0

ONEF: total 4 letters

by replacing plus and minus with on and off which mean the same.

Ali Humayun
  • 215
  • 1
  • 3
5

If we are allowed any encoding we like, we only need one letter. We choose some way of enumerating integers (say, in the order 0, 1, -1, 2, -2, 3, -3 ...) and then we encode the $n^\mathrm{th}$ integer as $n$ repetitions of our letter.

I don't think this is in the spirit of the question, which means that many of the other answers also aren't in the spirit of the question ("spell any integer in English"; I interpret this to mean, if you need to explain what number you are spelling, you have already lost). But if you're going to give an unspirited answer, you might as well give the optimal one :)

Ben Millwood
  • 847
  • 6
  • 10
3

The rules to this seem a little ambiguous but does this count? 4 letters:

{O,N,E,G}

Where all positive integers are just a that many ones:
1: one
2: one one
3: one one one


And negative integers are indicated by "neg" (an English abbreviation):
-1: neg one
-2: neg one one
-3: neg one one one


And zero is just an o:
0: o

Jason C
  • 240
  • 1
  • 8
3

5 (O N E L S)

For example: 110=ONE ONE ONE LESS ONE=111-1 Just do the lowest integer above your target value consisting of only ones, then subtract down. Works for any positive or negative number.

Ethan
  • 190
  • 5
1

2 answer depends on the rule

write numbers in binary representation (9 letters o,n,e,z,r,m,i,u,s) 1 = one
-1 = minus one
0 = zero
2 = one zero

or

or maybe just 6 letters (m,i,n,u,s,o) if we can write 1 as i and 0 as o (maybe invalid) 1 = i
-1 = minus i
0 = o
2 = io

Jamal Senjaya
  • 17,864
  • 2
  • 41
  • 147
1

We can refer to the number 1 using the less common word UNIT. Using MINUS adds two letters. Then using the word SUM in place of PLUS adds no new letters. Unlike PLUS, which is an operation between two operands, SUM naturally operates on all following operands.

-3: MINUS SUM UNIT UNIT UNIT
-2: MINUS SUM UNIT UNIT
-1: MINUS UNIT
 0: UNIT MINUS UNIT
 1: UNIT
 2: SUM UNIT UNIT
 3: SUM UNIT UNIT UNIT

If you're unhappy with the SUM operator taking an arbitrary number of operands, we can have it apply to only two operands. I've summarised below, using Polish notation for both SUM x y=x+y and MINUS x y=x-y.

-3: MINUS UNIT SUM SUM SUM UNIT UNIT UNIT UNIT
-2: MINUS UNIT SUM SUM UNIT UNIT UNIT
-1: MINUS UNIT SUM UNIT UNIT
 0: MINUS UNIT UNIT
 1: UNIT
 2: SUM UNIT UNIT
 3: SUM SUM UNIT UNIT UNIT

Thus we have 6 unique letters: IMNSTU

Samthere
  • 2,844
  • 13
  • 29
0

Five letters: {Z, E, R, O, N}

No maths required, but you have to understand binary. We spell out the bits involved in a binary representation of the desired integer. The first bit is the sign bit (ZERO for positive, ONE for negative).

Examples:

-3: ONE ONE ONE
-2: ONE ONE ZERO
-1: ONE ONE
 0: ZERO ZERO
 1: ZERO ONE 
 2: ZERO ONE ZERO
 3: ZERO ONE ONE
aroth
  • 129
  • 4
0

I used only SUB instead of PLUS and MINUS.

Using only six letters (B, E, N, O, U, S):

-3 : SUB (ONE ONE ONE)

-2 : SUB (ONE ONE)

-1 : SUB (ONE ONE)

0 : (ONE SUB ONE)

1 : SUB (SUB ONE)

2 : SUB (SUB ONE SUB ONE)

3 : SUB (SUB ONE SUB ONE)

  • Your answer relies on additional punctuation to avoid ambiguity, which feels like it's outside of the scope of good answers. Not a bad idea but it relies on interpreting or encoding language as well; we can obviously recognise the intended meaning of "sub" but it isn't standard. – Samthere Sep 12 '16 at 12:29
-1

Looking at all the number in binary give use the word zero and One which results in 5 characters.

8 = 1000 = one zero zero zero.

15 = 1111 = one one one one.

Rudy
  • 1
  • 1
    How is your answer any different from the ones already posted here before you? In addition you don't deal with the negative cases. – Shuri2060 Sep 12 '16 at 13:18