Independence and L ́evy Processes in Quantum Probability 93
φ 1 φ 2 (a 1 a 2 ···am) =
m
∏
k= 1
φek(ak),
φ 1 .φ 2 (a 1 a 2 ···am) = φ 1
(
→
∏
k:ek= 1
ak
)
∏
k:ek= 2
φ 2 (ak),
φ 1 /φ 2 (a 1 a 2 ···am) = ∏
k:ek= 1
φ 1 (ak)φ 2
( →
∏
k:ek= 2
ak
)
,
forφ 1 :A 1 →Candφ 2 :A 2 →Cand a typical elementa 1 a 2 ···am
∈ A 1 ‰A 2 ,ak∈ Aek,e 16 =e 26 =··· 6=em, that is, neighboringa’s
don’t belong to the same algebra. Note that we again have
canonical inclusionsAi → (A 1 ‰A 2 ),i = 1, 2, since the free
product without units is the coproduct in the category of not
necessarily unital algebras.
The monotone and anti-monotone product are not commutative,
but related by
φ 1 .φ 2 = (φ 2 /φ 1 )◦γA 1 ,A 2 ,
for all linear functionalsφ 1 :A 1 →C,φ 2 :A 2 →C, whereγA 1 ,A 2 :
A 1 ‰A 2 → A 2 ‰A 1 is the commutativity constraint (for the
commutativity constraint for the free product of unital algebras
see Equation (1.8.1)). The boolean product is commutative, that is,
it satisfies
φ 1 φ 2 = (φ 2 φ 1 )◦γA 1 ,A 2 ,
for all linear functionalsφ 1 :A 1 →C,φ 2 :A 2 →C.
Exercise 1.8.14 The boolean, the monotone, and the
anti-monotone products can also be defined for unital algebras, if
they are in the range of the unitization functor introduced in
Exercise 1.8.5.
Letφ 1 :A 1 →Candφ 2 :A 2 →Cbe two unital functionals on
algebrasA 1 ,A 2 , which can be decomposed asA 1 = C 1 ⊕A^01 ,
A 2 = C 1 ⊕ A^02. Then we define the boolean, monotone, or
anti-monotone product ofφ 1 andφ 2 as the unital extension of the
boolean, monotone, or anti-monotone product of their restrictions
φ 1 |A 0
1
andφ 2 |A 0
2
.