988 CHAPTER 27. ORLITZ SPACES

Defining Lv for v ∈ Ã by

Lv (u)≡∫

uvdµ,

it follows Lv ∈ LA (Ω)′ . From now on assume the measure space is σ finite. That is, thereexist measurable sets, Ωk satisfying the following:

Ω = ∪∞k=1Ωk, µ (Ωk)< ∞, Ωk ⊆Ωk+1.

Then

Proposition 27.2.1 For v ∈ LÃ (Ω) , the following inequality holds.

||v||Ã ≤ ||Lv|| ≤ 2 ||v||Ã .

Here Lv is considered as either an element of EA (Ω)′ or LA (Ω)′ and ||Lv|| refers to theoperator norm in either dual space.

Proof: The inequality 27.2.14 implies ||Lv|| ≤ 2 ||v||Ã . It remains to show the other halfof the inequality. If Lv = 0 there is nothing to show because this would imply that v = 0 soassume ||Lv||> 0. Define a measurable function, u, as follows. Letting r ∈ (0,1) ,

u(x)≡

{Ã(

r|v(x)|||Lv||

)/ v(x)||Lv|| if v(x) ̸= 0

0 if v(x) = 0.(27.2.15)

Now letFn ≡ {x : |u(x)| ≤ n}∩Ωn∩{x : v(x) ̸= 0} (27.2.16)

and defineun (x)≡ u(x)XFn (x) . (27.2.17)

Thus un is bounded and equals zero off a set which has finite measure. It follows that

A(|un|α

)∈ L1 (Ω)

for all α > 0. I claim that ||un||A ≤ 1. If not, there exists ε > 0 such that ||un||A− ε > 1.Then since A is convex,

1 <∫

A(

|un|||un||A− ε

)dµ ≤ 1

||un||A− ε

∫Ω

A(|un|)dµ.

Taking ε → 0+, using 27.1.6, and convexity of A along with 27.2.15 and 27.2.17,

||un||A ≤∫

A(|un|)dµ =∫

Fn

A(

rÃ(

r |v(x)|||Lv||

)/

rv(x)||Lv||

)dµ

≤ r∫

Fn

A(

Ã(

r |v(x)|||Lv||

)/

rv(x)||Lv||

)dµ ≤ r

∫Fn

Ã(

r |v(x)|||Lv||

)dµ

= r1||Lv||

∫Ω

un (x)v(x)dµ ≤ r1||Lv||

||un||A ||Lv||= r ||un||A ,

988 CHAPTER 27. ORLITZ SPACESDefining L, for v € A byLy(u)= [ wau,it follows L, € La (Q)’. From now on assume the measure space is o finite. That is, thereexist measurable sets, Q, satisfying the following:Q = UE Qg, W (Qe) < 09, Qe CQ.ThenProposition 27.2.1 For v € Lj (Q), the following inequality holds.Mla Sol] <2 Ivllz-Here Ly, is considered as either an element of E,(Q)' or La (Q)' and ||Ly|| refers to theoperator norm in either dual space.Proof: The inequality 27.2.14 implies ||Z,|| < 2]||v||q. It remains to show the other halfof the inequality. If Z,, = 0 there is nothing to show because this would imply that v = 0 soassume ||L,|| > 0. Define a measurable function, wu, as follows. Letting r € (0,1),A (PQ) ;2@) |u(x) = A( 4a) ie itv) £0 (27.2.15)O if v(x) =0.Now letF, = {x: |u(x)| <n}NQ,N {x: v(x) £0} (27.2.16)and defineUn (x) = u(x) ZF, (x). (27.2.17)Thus uw, is bounded and equals zero off a set which has finite measure. It follows thata( He) EL! (Q)for all @ > 0. I claim that ||u,||, <1. If not, there exists € > 0 such that ||u,||, —€ > 1.Then since A is convex,: |un| ) 1i< [a( dus —— | A Un|) dilJo Tally ae) “YS Tig we Jes’ elTaking € — 0+, using 27.1.6, and convexity of A along with 27.2.15 and 27.2.17,lenlla << fAlluade = fA G Cra )! Ta)~ (r|v(x)|\ rv(x) rly @)|s fa (Ga) qos faa)_ 7 [une)vean sr Pla sll = rll