Monad laws: Difference between revisions

From HaskellWiki
(Why is it an "associative law"?)
m (Remove leftover semicolon from previous edit)
 
(22 intermediate revisions by 9 users not shown)
Line 1: Line 1:
All instances of the [[Monad]] class should obey:
== The three laws ==


# "Left identity": <hask> return a >>= f  =  f a </hask>
All instances of the [[Monad]] typeclass should satisfy the following laws:
# "Right identity": <hask> m >>= return  =  m </hask>
# "Associativity":  <hask> (m >>= f) >>= g = m >>= (\x -> f x >>= g) </hask>


== What is the practical meaning of the monad laws? ==
{|
|-
| ''Left identity:''
| <haskell>return a</haskell>
| <tt>'''>>='''</tt>
| <haskell>h</haskell>
| ≡
| <haskell>h a</haskell>
|-
| ''Right identity:''
| <haskell>m</haskell>
| <tt>'''>>='''</tt>
| <haskell>return</haskell>
| ≡
| <haskell>m</haskell>
|-
| ''Associativity:''
| <haskell>(m >>= g)</haskell>
| <tt>'''>>='''</tt>
| <haskell>h</haskell>
| ≡
| <haskell>m >>= (\x -> g x >>= h)</haskell>
|}


Let us re-write the laws in do-notation:
Here, ''p'' ≡ ''q'' simply means that you can replace ''p'' with ''q'' and vice-versa, and the behaviour of your program will not change: ''p'' and ''q'' are equivalent.


<haskell>
Using eta-expansion, the ''associativity'' law can be re-written for clarity as:
1. do { x' <- return x           do { f x
 
      ; f x'              ==        }
{|
      }
|
|<haskell>(m  >>=  (\x -> g x))  >>=  h</haskell>
|-
| ≡
|<haskell> m  >>=  (\x -> g x   >>=  h)</haskell>
|}
 
or equally:
 
{|
|
|<haskell>(m  >>=  (\x -> g x))  >>=  (\y -> h y)</haskell>
|-
| ≡
|<haskell> m  >>=  (\x -> g x   >>= (\y -> h y))</haskell>
|}


2. do { x <- m            ==     do { m
=== Is that really an "associative law"? ===
      ; return x }                  }
----
In this form, not at first glance. To see precisely why they're known as ''identity'' and ''associative'' laws, you have to change your notation slightly.


3. do { y <- do { x <- m          do { x <- m
The monad-composition operator <code>(>=>)</code> (also known as the ''Kleisli-composition'' operator) is defined in <code>Control.Monad</code>:
                ; f x                ; do { y <- f x
                }          ==            ; g y
      ; g y                              }
      }                              }


                                  do { x <- m
<haskell>
                    using 3.14      ; y <- f x
infixr 1 >=>
                          ==        ; g y
(>=>)  :: Monad m => (a -> m b) -> (b -> m c) -> (a -> m c)
                                    }
f >=> g =  \x -> f x >>= g
</haskell>
</haskell>


In this notation the laws appear as plain common sense.
Using this operator, the three laws can be expressed like this:


== But why should monads obey these laws? ==
{|
|-
| ''Left identity:''
| <haskell>return</haskell>
| <tt>'''>=>'''</tt>
| <haskell>h</haskell>
| ≡
| <haskell>h</haskell>
|-
| ''Right identity:''
| <haskell>f</haskell>
| <tt>'''>=>'''</tt>
| <haskell>return</haskell>
| ≡
| <haskell>f</haskell>
|-
| ''Associativity:''
| <haskell>(f >=> g)</haskell>
| <tt>'''>=>'''</tt>
| <haskell>h</haskell>
| ≡
| <haskell>f >=> (g >=> h)</haskell>
|}


When we see a program written in a form on the LHS, we expect it to do
It's now easy to see that monad composition is an associative operator with left and right identities.
the same thing as the corresponding RHS; and vice versa.  And in
practice, people do write like the lengthier LHS once in a while.
First example: beginners tend to write


<haskell>
This is a very important way to express the three monad laws, because they are precisely the laws that are required for monads to form a mathematical [[Category theory|category]].  To summarise in ''haiku'':
skip_and_get = do { unused <- getLine
                  ; line <- getLine
                  ; return line
                  }
</haskell>


and it would really throw off both beginners and veterans if that did
:<tt>
not act like (by law #2)
:''Monad axioms:''
:''Kleisli composition forms''
:''a category.''
:</tt>


<haskell>
== The monad laws in practice ==
skip_and_get = do { unused <- getLine
                  ; getLine
                  }
</haskell>


Second example: Next, you go ahead to use skip_and_get:
If we re-write the laws using Haskell's <code>do</code>-notation:


<haskell>
{|
main = do { answer <- skip_and_get
|-
          ; putStrLn answer
| ''Left identity:''
          }
| <haskell>
do
  x′ <- return x
  f x′
</haskell>
| ≡
| <haskell>
do
  f x
</haskell>
|-
| ''Right identity:''
| <haskell>
do
  x <- m
  return x
</haskell>
| ≡
| <haskell>
do
  m
</haskell>
|-
| ''Associativity:''
| <haskell>
do
  y <- do
    x <- m
    f x
  g y
</haskell>
| ≡
| <haskell>
do
  x <- m
  do
    y <- f x
    g y
</haskell>
| ≡
| <haskell>
do
  x <- m
  y <- f x
  g y
</haskell>
</haskell>
|}


The most popular way of comprehending this program is by inlining
we can see that the laws represent plain, ordinary common-sense transformations of imperative programs.
(whether the compiler does or not is an orthogonal issue):


<haskell>
=== But why should monadic types satisfy these laws? ===
main = do { answer <- do { unused <- getLine
----
                        ; getLine
When we see a program written in a form on the left-hand side, we expect it to do the same thing as the corresponding right-hand side; and vice versa.  And in practice, people do write like the lengthier left-hand side once in a while.
                        }
          ; putStrLn answer
          }
</haskell>


and applying law #3 so you can pretend it is
* First example: beginners tend to write


<haskell>
:<haskell>
main = do { unused <- getLine
skip_and_get = do
          ; answer <- getLine
  unused <- getLine
          ; putStrLn answer
  line <- getLine
          }
  return line
</haskell>
</haskell>


Law #3 is amazingly pervasive: you have always assumed it, and you
:and it would really throw off both beginners and veterans if that did not act like (by ''right identity''):
have never noticed it.


Whether compilers exploit the laws or not, you still want the laws for
:<haskell>
your own sake, just so you can avoid pulling your hair for
skip_and_get = do
counter-intuitive program behaviour that brittlely depends on how many
  unused <- getLine
redundant "return"s you insert or how you nest your do-blocks.
  getLine
</haskell>


== But it doesn't look exactly like an "associative law"... ==
* Second example: Next, you go ahead and use <code>skip_and_get</code>:


Not in this form, no.  To see precisely why they're called "identity laws" and an "associative law", you have to change your notation slightly.
:<haskell>
main = do
  answer <- skip_and_get
  putStrLn answer
</haskell>


The monad composition operator (also known as the Kleisli composition operator) is defined in Control.Monad:
:The most popular way of comprehending this program is by ''inlining'' (whether the compiler does or not is an orthogonal issue):


<haskell>
:<haskell>
(>=>) :: Monad m => (a -> m b) -> (b -> m c) -> a -> m c
main = do
m >=> n = \x -> do { y <- m x; n y }
  answer <- do
    unused <- getLine
    getLine
  putStrLn answer
</haskell>
</haskell>


Using this operator, the three laws can be expressed like this:
:and applying associativity so you can pretend it is:


# "Left identity": <hask> return >=> g = g </hask>
:<haskell>
# "Right identity": <hask> f >=> return = f </hask>
main = do
# "Associativity":  <hask> (f >=> g) >=> h = f >=> (g >=> h) </hask>
  unused <- getLine
  answer <- getLine
  putStrLn answer
</haskell>


It's now easy to see that monad composition is an associative operator with left and right identities.
The associativity law is amazingly pervasive: you have always assumed it, and you have never noticed it.  


This is a very important way to express the three monad laws, because they are precisely the laws that are required for monads to form a mathematical [[Category theory|category]].  So the monad laws can be summarised in convenient Haiku form:
The associativity of a ''binary'' operator allows for any number of operands to be combined by applying the binary operator with any arbitrary grouping to get the same well-defined result, just like the result of summing up a list of numbers is fully defined by the binary <code>(+)</code> operator no matter which parenthesization is used (yes, just like in folding a list of any type of monoidal values).


<blockquote>
Whether compilers make use of them or not, you still want the laws for your own sake, just so you can avoid pulling your hair out over counter-intuitive program behaviour, which depends (in brittle fashion!) on e.g. how many redundant <code>return</code>s you insert or how you nest your <code>do</code>-blocks...
Monad axioms:<br/>
Kleisli composition forms<br/>
a category.
</blockquote>


[[Category:Standard_classes]]
[[Category:Standard_classes]]
[[Category:Monad]]
[[Category:Monad]]

Latest revision as of 10:40, 1 February 2024

The three laws

All instances of the Monad typeclass should satisfy the following laws:

Left identity:
return a
>>=
h
h a
Right identity:
m
>>=
return
m
Associativity:
(m >>= g)
>>=
h
m >>= (\x -> g x >>= h)

Here, pq simply means that you can replace p with q and vice-versa, and the behaviour of your program will not change: p and q are equivalent.

Using eta-expansion, the associativity law can be re-written for clarity as:

(m  >>=  (\x -> g x))  >>=  h
 m  >>=  (\x -> g x    >>=  h)

or equally:

(m  >>=  (\x -> g x))  >>=  (\y -> h y)
 m  >>=  (\x -> g x    >>=  (\y -> h y))

Is that really an "associative law"?


In this form, not at first glance. To see precisely why they're known as identity and associative laws, you have to change your notation slightly.

The monad-composition operator (>=>) (also known as the Kleisli-composition operator) is defined in Control.Monad:

infixr 1 >=>
(>=>)   :: Monad m => (a -> m b) -> (b -> m c) -> (a -> m c)
f >=> g =  \x -> f x >>= g

Using this operator, the three laws can be expressed like this:

Left identity:
return
>=>
h
h
Right identity:
f
>=>
return
f
Associativity:
(f >=> g)
>=>
h
f >=> (g >=> h)

It's now easy to see that monad composition is an associative operator with left and right identities.

This is a very important way to express the three monad laws, because they are precisely the laws that are required for monads to form a mathematical category. To summarise in haiku:

Monad axioms:
Kleisli composition forms
a category.

The monad laws in practice

If we re-write the laws using Haskell's do-notation:

Left identity:
do
  x′ <- return x
  f x′
do
  f x
Right identity:
do
  x <- m
  return x
do
  m
Associativity:
do
  y <- do
    x <- m
    f x
  g y
do
  x <- m
  do
    y <- f x
    g y
do
  x <- m
  y <- f x
  g y

we can see that the laws represent plain, ordinary common-sense transformations of imperative programs.

But why should monadic types satisfy these laws?


When we see a program written in a form on the left-hand side, we expect it to do the same thing as the corresponding right-hand side; and vice versa. And in practice, people do write like the lengthier left-hand side once in a while.

  • First example: beginners tend to write
skip_and_get = do
  unused <- getLine
  line <- getLine
  return line
and it would really throw off both beginners and veterans if that did not act like (by right identity):
skip_and_get = do
  unused <- getLine
  getLine
  • Second example: Next, you go ahead and use skip_and_get:
main = do
  answer <- skip_and_get
  putStrLn answer
The most popular way of comprehending this program is by inlining (whether the compiler does or not is an orthogonal issue):
main = do
  answer <- do
    unused <- getLine
    getLine
  putStrLn answer
and applying associativity so you can pretend it is:
main = do
  unused <- getLine
  answer <- getLine
  putStrLn answer

The associativity law is amazingly pervasive: you have always assumed it, and you have never noticed it.

The associativity of a binary operator allows for any number of operands to be combined by applying the binary operator with any arbitrary grouping to get the same well-defined result, just like the result of summing up a list of numbers is fully defined by the binary (+) operator no matter which parenthesization is used (yes, just like in folding a list of any type of monoidal values).

Whether compilers make use of them or not, you still want the laws for your own sake, just so you can avoid pulling your hair out over counter-intuitive program behaviour, which depends (in brittle fashion!) on e.g. how many redundant returns you insert or how you nest your do-blocks...