※ このテキストは、以下のプレプリントの前半部分を、個人的な学習目的のため和約したものです。 ※ arXiv.org > quant-ph > arXiv:quant-ph/0210044 ※ http://arxiv.org/abs/quant-ph/0210044 Physical content of Heisenberg’s uncertainty relation: Limitation and reformulation ハイゼンベルグの不確定性関係の物理的内容:限界と再定式化 Heisenberg’s reciprocal relation between position measurement error and momentum disturbance is 位置の観測誤差と、運動量の撹乱の間に成り立つハイゼンベルグの相反関係は、 rigorously proven under the assumption that those error and disturbance are independent of the state of the measured object. 厳密には、それら観測対象の状態の誤差と撹乱が独立であるという仮定の下に証明されたものである。 A generalization of Heisenberg’s relation proven valid for arbitrary measurements is proposed 一般化された任意の観測について有効性が示されるハイゼンベルグの関係と、 and reveals two distinct types of possible violations of Heisenbergs relation. 2つの際立ったタイプのハイゼンベルグの関係の生じ得る破れについて提案しよう。 -------------------------------------------------------------------------------- I. INTRODUCTION An essential departure of quantum mechanics from classical mechanics is 量子力学と古典力学の本質的な決別は、 that any measurement of a microscopic object involves the interaction with the apparatus not to be neglected いかなるミクロな対象の観測も、無視することのできない観測機器との相互作用を含み、 and accordingly introduces an unavoidable and uncontrollable disturbance on the measured object. それゆえ、観測対象の物体の上に不可避でコントロールできない撹乱を生み出すということだ。 Undoubtedly, this point of view led to fundamental doctrines of the Copenhagen interpretation of quantum mechanics [1], 疑いもなく、量子力学のコペンハーゲン解釈の基本的な教義はこの視点から導かれたものであり、 which successfully dissolved the wave-particle duality and the continuous-discontinuous discrepancy. そしてそれは波と粒子の二重性や連続-非連続の食い違いを見事に解消した。 In his celebrated paper [2] published in 1927, Heisenberg attempted to establish the quantitative expression of the amount of unavoidable momentum disturbance caused by any position measurement. ハイゼンベルグは、有名な1927年の論文の中で、位置の観測によって引き起こされる不可避な運動量の撹乱についての定式化を試みた。 His statement, with some elaborations, can be formulated as follows: 彼の主張は、多少の技巧と共に、次のように定式化できるだろう: For every measurement of the position Q of a mass with root-mean-square error ε(Q), 物体の位置Qの観測誤差の二乗平均平方根を ε(Q) とすれば、 the root-mean-square disturbance η(P) of the momentum P of the mass caused by the interaction of this measurement 観測によって引き起こされる物体の運動量Pの二乗平均平方根 η(P) は、 always satisfies the relation 常に次の関係を満たす ε(Q)η(P) >= h~ / 2 ・・・(1) Heisenberg [2] not only explained the physical intuition underlying the above relation by discussing the γ ray microscope thought experiment, ハイゼンベルグはガンマ線顕微鏡の試行実験によって、上記関係の基礎となる物理的な直観を説明するのみならず、 but also claimed that this relation can be proven as a straightforward consequence of the canonical commutation relation (CCR), QP - PQ = i h~ この関係は正準交換関係 QP - PQ = i h~ からの直接の帰結としても示されると主張している。 A mathematical part of his proof was refined by introducing the notion of standard deviation shortly afterward by Kennard [3] 証明の数学的な部分は、カナード以降、標準偏差の概念の導入により洗練され、 and later generalized to arbitrary pair of observables by Robertson [4] as the following statement: 後にロバートソンによって任意のオブザーバブル(物理的観測量)の組へと一般化された、それは次のようなものだ: For any pair of observables A and B, 任意のオブザーバブル A, B の組について、 their standard deviations, σ(A) and σ(B), satisfy the relation それらの標準偏差 σ(A) と σ(B) は、次の関係を満たす σ(A)σ(B) >= 1/2 [ < ψ, [A,B] ψ > ] ・・・(2) in any state ψ with σ(A), σ(B) < ∞. いかなる状態 ψ、σ(A), σ(B) < ∞ についても。 In the above, [A,B] stands for the commutator [A,B] = AB - BA, 上で、[A,B] とは 交換子 [A,B] = AB - BA のことであり、 and the standard deviation is defined as σ(A) = (<ψ, A^2 ψ> - <ψ, Aψ>^2 ) ^ 1/2, 標準偏差は次のように定義される σ(A) = (<ψ, A^2 ψ> - <ψ, Aψ>^2 ) ^ 1/2 where <・・・, ・・・> denotes the inner product [5]. ここで <・・・, ・・・> は内積を表す。 As a consequence of [Q, P] = ih~, we have [Q, P] = ih~ の結果として、次の式を得る σ(Q)σ(P) >= h~ / 2 ・・・(3) which was proven by Heisenberg himself for Gaussian states and by Kennard [3] generally. この式はハイゼンベルグ自身によってガウス分布的な状態について証明し、後にカナードが一般的な状態について証明した。 Heisenberg [2] argued that the mathematical relation Eq. (3) concludes the physical assertion expressed by Eq. (1). ハイゼンベルグは、数学的な関係式(3)は、(1) 式に示される物理的な表明の帰結であると述べている。 Since then, his claim has been accepted by many [6, 7, 8, 9, 10, 11, 12]. それ以来、彼の主張は多くに受け入れられてきた。 However, the universal validity of Eq. (1) has been also criticized in many ways [13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29]. しかし、式(1) の不変的な正当性については、また多方面からの批判があった。 In fact, the “resolution power” of the γ ray microscope cannot be identified in any interpretation 実際、ガンマ線顕微鏡の「分解能」は、いかなる解釈においても、 with the standard deviation of the mass position in the state to be measured. 観測された状態での物体の位置の標準偏差によって定められない。 Thus, Eq. (3) cannot be considered as a “formal expression” of Eq. (1). 従って、式(3) は 式(1) の”正式な表現”だと見なすわけにはゆかない。 Moreover, no one has succeeded in proving Eq. (1) for general measurements even by applying Eq. (3) not only to the mass state but also to the apparatus state. さらに、一般的な観測について、式(3)を物体の状態だけでなく観測機器の状態にまで適用して、式(1) を証明した者はまだ誰もいない。 Undoubtedly, this has caused serious confusions among physicists on the status of this leading principle of quantum mechanics. 疑いもなくこのことは、量子力学の主要な原理の地位を巡って、物理学者の間に深刻な混乱を引き起こしている。 In order to clarify the confusion, この混乱を明らかにするため、 we shall start with precise definitions of root-mean-square disturbance and mean-square error in arbitrary measurement 我々は、任意の観測における撹乱の二乗平均平方と誤差の二乗平均を明確に定義することから始め、 and reconstruct Heisenberg’s argument. ハイゼンベルグの議論を再構築しよう。 His argument applies Eq. (3) to the mass state just after the measurement. 彼の議論では、式(3) を観測直後の物体の状態に当てはめている。 However, in order to derive Eq. (1), additional assumptions are necessary. しかし、式(1)の導出のためには、さらなる仮定が必要となる。 Here, we shall show that the following two assumptions completes Heisenberg’s argument: そこで我々は、ハイゼンベルグの議論を完備するため、次なる2つの仮定を示すべきだろう: (i) Both amounts of ε(Q) and η(P) are independent of the input state. (i) 入力状態における2つの量、ε(Q) と η(P) は互いに独立である。 (ii) The measurement always leaves the mass with position standard deviation smaller than the ε(Q). (ii) 観測はいつでも物体を ε(Q) よりも小さな標準偏差の位置に置いたままにする。 Thus, Heisenberg’s argument left open the following questions: それゆえ、ハイゼンベルグの議論は次のような疑問を残したままである: Can we further relax the assumptions? 我々は、さらに仮定を緩めることができるか? What relation holds for arbitrary measurements? 任意の観測について、いかなる関係が成り立ったままなのか? What conditions characterize violations of Eq. (1)? どんな条件が、式(1)の破れを特徴付けるのか? In particular, the second assumption, which will be called the equipredictivity of measurement, 特に、2番目の仮定 〜これを観測の等価予測性(equipredictivity)と呼ぼう〜 stringently restricts the class of measurements to which Eq. (1) is applicable. は、式(1)が適用できる観測の種類を厳しく制限する。 In order to answer those questions, a new approach to proving Eq. (1) will be proposed これらの疑問に答えるため、式(1)を証明する新たなアプローチが望まれる、 based on analysis of commutation relations for the operators representing the error and the disturbance. 誤差と撹乱に代表される演算子についての交換関係の解析に基づいた。 From this approach, we shall obtain a generalization of Heisenberg’s relation proven valid for arbitrary measurements このアプローチから我々は、任意の観測について一般化したハイゼンベルグの関係の有用性を示し、 and also obtained a new proof of Heisenberg’s relation without assuming the equipredictivity. そしてまた、等価予測性の仮定無しでハイゼンベルグの関係の新たな証明を行おう。 We shall further discuss limitations on Heisenberg’s relation based on the generalized relation さらに我々は、一般化された関係に基づいたハイゼンベルグの関係の限界について議論し、 and show that there are two distinct types of measurements in which Heisenberg’s relation is violated uniformly for any input state. ハイゼンベルグの関係がいかなる入力状態に対しても一様に破れている2つの際立ったタイプがあることを示そう。 -------------------------------------------------------------------------------- II. FORMULATION OF ERROR AND DISTURBANCE Let us consider a measurement of position Q of a mass with momentum P. 運動量Pの物体の、位置Qを観測することを考えてみよう。 The interaction between the mass and the apparatus is assumed to turn on at time 0 and turn off at timeΔt. 物体と観測機器の相互作用は、時刻0に始まり、時刻Δtで終わるものだとしよう。 Let ψ be the input state, i.e., the state of the mass just before the interaction, ψを入力状態、つまり、相互作用が始まる直前の状態として、 and let ξ be the state of the apparatus just before the interaction. ξを観測機器の相互作用が始まる直前の状態だとしよう。 Let U be the unitary operator representing the time evolution of the mass plus apparatus for the time interval (0,Δt).時刻0からΔtの間の、物体と観測機器の時間発展を表すユニタリー演算子をUとしよう。 Then, in the Heisenberg picture the momentum change is represented by ここで、ハイゼンベルグ描像での運動量の変化は次のように表せる、 D(P) = P(Δt) - P(0) ・・・(4) where P(0) = P (x) I and P(Δt) = U†P(0)U. ここで P(0) = P (x) I 、   ※ (x)はテンソル積、もともとは○の中にxを入れた記号。 P(Δt) = U†P(0)U である。   ※ P(0) のUによる時間発展。 The root-mean-square momentum disturbance is defined by 運動量の撹乱の二乗平均平方根は、次のように定義される、 η(P) = < D(P)^2 >^1/2 ・・・(5) where <・・・> denotes the mean value in the original state ψ (x) ξ. ここで <・・・> は、元の状態 ψ (x) ξ の平均を表す。 Now we shall discuss some basic properties of the above quantity. さて、我々は以上の量の基本的な性質について議論を進めよう。 The above statistical definition leads to the following geometric expression 上の統計的な定義から、次のような幾何学的な表現が導かれる、 η(P) = || P(Δt)ψ(x)ξ - P(0)ψ(x)ξ || ・・・(6) Thus, by the relation ここで、任意のオブザーバブルAに対して成り立つ関係 σ(A) = || A ψ(x)ξ - ψ(x)ξ || ・・・(7) for any observable A, we easily obtain the following relation より、我々は簡単に次の関係式を得る、 |σ(P(Δt)) - σ(P(0))| <= η(P) + | - | ・・・(8) stating that the change in the momentum standard deviation is bounded from above by the root-mean-square disturbance except for the change in the mean momentum. 運動量の標準偏差の変化の大きさは、運動量の平均の変化を除いた撹乱の二乗平均平方根によって制限される、ということだ。 If the mass had a definite momentum p before the measurement, i.e., P(0)ψ = pψ, we would have もし物体が観測の前にはっきりした運動量pを持っていれば、つまり、P(0)ψ = pψならば、次の式を得る、 D(P)ψ(x)ξ = [P(Δt) - p]ψ(x)ξ ・・・(9) and hence 従って η(P)^2 = < [P(Δt) - p]^2 > >= σ(P(Δt))^2 ・・・(10) Thus, the momentum standard deviation after the measurement is bounded from above by the root-mean-square disturbance. 故に、観測後の運動量の標準偏差は、撹乱の二乗平均平方根によって制限される。 In particular, if p = 0, we have 特に、もし p = 0 ならば、次のようになる、 η(P)^2 = < P(Δt)^2 > ・・・(11) This relation can be interpreted as follows. この関係は、次のように解釈できるだろう、 If the mass is at rest before the measurement, もし物体が観測以前に静止していれば、 all the object’s momentum after the measurement arises from the interaction, 相互作用から引き起こされる観測後の全ての物体の運動量は、 so that the mean-square disturbance (i.e., the square of the root-mean-square disturbance) 攪乱の二乗平均(つまり、撹乱の二乗平均平方根の二乗) is equal to the mean-square momentum after the measurement. が観測後の運動量の二乗平均に等しくなるようになっている。 The role of the interaction is to transduce the value of Q before the interaction 相互作用の役目は、相互作用以前の位置Qの値を、 to the value of an observable M of the probe, a part of the apparatus, after the interaction. 相互作用以後の、観測装置の一部である探査機のオブザーバブルMの値に変換することである。 We shall call M the meter observable. 我々はMのことをメーターオブザーバブルと呼ぼう。 We suppose that 我々は次のように考えている、 after the interaction is turned off, the outcome of the Q measurement in the state ψ is 相互作用が終わった後、状態ψにおけるQの観測結果は、 obtained by measuring M without further disturbing the momentum P of the mass; Mを測ることによって、もうそれ以上物体の運動量Pを乱すことなしに得られる、 this is possible by another measuring apparatus coupled only to the probe. そのような観測は、探査機のみと結びついた他の観測装置によって、可能である。 The postulates of quantum mechanics do not limit the accuracy of the latter measurement of M, 量子力学の前提条件では、後からMを観測する精度を制限しているわけではないので、 and hence we neglect the error from this measurement. 我々はこの観測につてのエラーを無視できる。 Then, in the Heisenberg picture the error caused by this process of the Q measurement is represented by それゆえハイゼンベルグ描像において、Qの観測過程で引き起こされる誤差は次のようになる、 N(Q) = M(Δt) - Q(0) ・・・(12) where Q(0) = Q (x) I and M(Δt) = U† (I (x) M) U. ここで Q(0) = Q (x) I であり、M(Δt) = U† (I (x) M) U である。 ※ (x) はテンソル積の記号 The root-mean-square error is defined by 誤差の二乗平均平方根の定義は、 ε(Q) = < N(Q)^2 > ^ 1/2 ・・・(13) The above statistical definition leads to the following geometric expression 上の統計的な定義から、以下の幾何学的な定義が導かれる ε(Q) = || M(Δt) ψ (x) ξ - Q(0) ψ (x) ξ || ・・・(14) Analogously with Eq. (15), we obtain 同様にして、次式(15)を得る、 |σ(M(Δt)) - σ(Q(0))| <= ε(Q) + | - | ・・・(15) showing that the change in the standard deviation from input to output is この式は次のことを示している、入力と出力の標準偏差の変化の上限は、 bounded from above by the root-means-quare noise except for the bias, namely, the change in the mean value. バイアス(つまり、平均値の変化)を除いた誤差の二乗平均平方によって抑えられる。 If the mass had a definite position q before the measurement, we would have もし物体の観測以前の位置qがはっきりしていれば、次のようになる、 σ(Q)^2 = <[M(Δt) - q]^2 >= σ(M(Δt))^2 ・・・(16) so that ε(Q) would represent the root-mean-square deviation of the measurement outcome from the true position q ε(Q)は、真の位置qの観測結果の撹乱の二乗平均平方根を表しており、 and would limit the standard deviation of the outcome of the Q measurement. それはQの結果の標準偏差を制限している。 -------------------------------------------------------------------------------- III. RECONSTRUCTION OF HEISENBERG’S ARGUMENT In the 1927 paper, Heisenberg claimed that Eq. (1) is a straightforward mathematical consequence of the CCR, QP - PQ = ih~. 1972の論文でハイゼンベルグは、式(1)を正準交換関係 QP - PQ = ih~ の数学的な直接の帰結であると論じた。 Although he gave a claimed proof of Eq. (1) from the CCR, たとえ彼が正準交換関係から式(1)の証明を与えたと言っていたとしても、 his argument includes some implicit assumptions. 彼の議論には隠れた幾つかの仮定が含まれていた。 Consequently, Eq. (1) cannot be considered as a straightforward consequence of the CCR 結論として、式(1)は正準交換関係の直接の帰結には成り得ないし、 nor as a universally valid relation. 普遍的に成り立つ関係でもない。 Heisenberg’s idea of the proof is as follows: ハイゼンベルグの証明のアイデアは次のようなものだ: Start with a measurement with error ε(Q), 観測の誤差 ε(Q) から始めて、 claim that the state just after the measurement has σ(Q) = ε(Q), 観測直後の状態が σ(Q) = ε(Q) であると論じ、 prove Eq. (3) from the CCR for the above state 上の状態の正準交換関係から式(3)を証明して to obtain the relation ε(Q)σ(P) >= h~/2, 関係式 ε(Q)σ(P) >= h~/2 を得て、 and identify σ(P) with η(P) to obtain Eq. (1). そして σ(P) を η(P) だと同定して式(1)を得る。 In order to clarify the hidden assumptions, in what follows, ハイゼンベルグの仮定をはっきりさせるため、次の中で、 we shall show that Heisenberg’s original proof can be rigorously reconstructed under the following two additional assumptions: 我々はハイゼンベルグのオリジナルの証明が、以下の2つの仮定を付け加えることによって、厳密に再構築されることを示そう。 (i) Both ε(Q) and η(P) are independent of the input state; (i) ε(Q) と η(P) は共に入力状態では独立である、 in this case, we say that the measurement has constant mean-square noise and disturbance. この場合、観測は一定の二乗平均誤差と撹乱を持つと言えるだろう。 (ii) The measurement always leaves the mass with position standard deviation smaller than ε(Q); 観測はいつでも物体を ε(Q) よりも小さな標準偏差の位置に置いたままにする、 in this case, we say that the measurement is equipredictive. この場合、観測は等価予測的(equipredictive)であると言えるだろう。 Under the above assumptions, the proof of Eq. (1) runs as follows. 上の仮定の下で、式(1)の証明を以下に行おう。 In order to obtain an estimate of the momentum disturbance, 運動量の撹乱の見積りを得るため、 we shall consider the case where the mass were at rest before the measurement. 物体が観測の前に静止している場合を考えてみよう。 Then, by Eq. (11) the object’s mean-square momentum after the measurement ここで式(11)によって、物体の測定後の二乗平均運動量は is equal to the mean-square momentum disturbance, i.e., 二乗平均運動量の撹乱に等しくなる、つまり、 η(P)^2 = < P(Δt)^2 > ・・・(17) Let ψx be the state of the mass after the measurement with outcome x. 物体の、測定後の結果 x だったときの状態を ψx としよう。 We shall denote by σx the standard deviation in the state ψx. 状態 ψx での標準偏差が σx であるとする。 We shall later show that the relation 我々は後に、次の関係式 η(P) >= σx(P) ・・・(18) holds with positive probability. が正の確率を持つことを示そう。 On the other hand, by condition (ii), we have 一方、仮定(ii)より、我々は ε(Q) >= σx(Q) ・・・(19) holds with probability one. が確率1で成り立つことを得る。 It follows from Eqs. (18) and (19) that there must be a state vector ψx satisfying 式(18)と(19)より、状態ベクトル ψx は次の式を満たさなければならない、 ε(Q)η(P) >= σx(Q)σx(P) ・・・(20) Thus, Eq. (1) follows from Eq. (3), if the mass is at rest just before the measurement. かくして、式(1)は 式(3) から得られることになる、物体が測定前に静止していれば。 Then, assumption (i) ensures that Eq. (1) holds irrespective of the particular choice of the input state. そして、仮定(i)によって、式(1)が入力状態の特定の選択とは関係なしに成り立つことが保障される。 We, therefore, conclude that そこで、我々は結論づける、 every equipredictive measurement with constant mean-square noise and disturbance satisfies Eq. (1) for every input state [31]. 一定の二乗平均誤差と撹乱をともなう、あらゆる等価予測的な測定は、あらゆる入力状態に対して式(1)を満たすと。 Now, we shall prove Eq. (18). さてここで、式(18)を証明しよう。 The state ψx arises with the probability distribution π(dx) of obtaining the outcome x determined by 状態 ψx は、次式で定義される、結果 x より得られる確率分布 π(dx) によって生じる。 π(dx) = < E^(M(t)) (dx) > ・・・(21) where E^(M(Δt)) is the resolution of the identity corresponding to the observable M(Δt). ここで E^(M(Δt)) は、対応するオブザーバブル M(Δt) を識別するための解像度である。 Then, since < P(Δt)^2 > is the mean of P^2 after the measurement, we have そして、< P(Δt)^2 > は観測後の p^2 の平均値であることから、 < P(Δt)^2 > = ∫ < ψx, P^2 ψx > π(dx) ・・・(22) Since < ψx, P^2 ψx > >= σx (P)^2 for any x, we have あらゆる x について < ψx, P^2 ψx > >= σx (P)^2 が成り立つことから、 ∫ < ψx, P^2 ψx > π(dx) >= ∫σx (P)^2 π(dx) ・・・(23) and consequently その結果として、 < P(Δt)^2> >= ∫ σx(P)^2 π(dx) ・・・(24) Thus, we have かくして、我々は < P(Δt)^2 > >= σx(P)^2 ・・・(25) with positive probability; を正の確率で得る。 otherwise the opposite inequality of Eq. (24) would hold. あるいは、反対の不等式(24)が保たれるだろう。 This concludes the proof of Eq. (18). 以上が式18)の証明だ。 -------------------------------------------------------------------------------- IV. UNIVERSALLY VALID REFORMULATION The above proof of Eq. (1) rigorously reproduces Heisenberg’s intention in 1927 上記の式(1)の証明は、1927年のハイゼンベルグの意図を厳密に再現したものだ、 that the uncertainty relation expressed by Eq. (1) is a straightforward consequence from Eq. (3) for the state after the measurement. そこでは、式(1)で示される不確定性関係が、観測後の状態についての式(3)からの直接の帰結となっている。 However, the above proof does not show that Eq. (1) is universally valid しかし、上記の証明は、式(1)が普遍的に成り立つことを示しておらず、 and has left the problem quite open as to the limitation of Eq. (1). 式(1)の制限について、かなりはっきりした問題を残している。 Thus, the following problems have remained open on Heisenberg’s uncertainty relation. すなわち、次のような問題がハイゼンベルグの不確定性関係の上に残されたままなのである。 Can we further relax the assumptions of the proof? 証明の仮定をもっと緩めることはできないだろうか? What relation holds for arbitrary measurements? 任意の観測について、いかなる関係が成り立つのか? What conditions characterize violations of Eq. (1)? どんな条件が式(1)の破れを特徴づけるのか? In what follows, we shall prove that the relation 続いて、我々は証明しよう、次の関係式 ε(Q)η(P) + ε(Q)σ(P) + σ(Q)η(P) <= h~/2 ・・・(26) ※これがいわゆる「小澤の不等式」 holds for every measurement and every input state as long as all the relevant terms are finite, が、関連する項が有限であるような任意の観測と入力状態について成立することを、 where σ(Q) and σ(P) refer to the standard deviations of Q and P in the input state. ここで σ(Q) と σ(P) は、入力状態の Q と P の標準偏差を示している。 The proof of Eq. (26) runs as follows. 式(26)の証明を以下に進める。 Since M and P are observables in different systems, we have [M(Δt), P(Δt)] = 0. M と P が異なる系のオブザーバブルであるなら、[M(Δt), P(Δt)] = 0 である。 Substituting M(Δt) = Q(0) + N(Q) and P(Δt) = P(0) + D(P) for this relation この関係の M(Δt) を Q(0) + N(Q) に、 P(Δt) を P(0) + D(P) に置き換えて、 and using the commutation relation [Q, P] = ih~, we have [18, 23, 24] [Q, P] = ih~ の関係を用いれば、次のようになる [N(Q),D(P)] + [N(Q), P(0)] + [Q(0),D(P)] = -ih~ ・・・(27) Taking the moduli of means of the both sides and applying the triangular inequality, we have 両辺の平均の係数をとって、三角不等式を適用すると、 |< [N(Q),D(P)] >| + |< [N(Q), P(0)] >| + |< [Q(0),D(P)] >| >= h~ ・・・(28) Since the variance is not greater than the mean-square, Eq. (2) gives 分散は二乗平均より大きくならないことから、式(2)は次の関係を与える ε(Q)η(P) >= σ[N(Q)]σ[D(P)] >= 1/2 |< [N(Q),D(P)] >| ・・・(29) where σ in the middle term refers to the standard deviation in the state ψ (x) ξ. ここで中央の項のσは、状態 ψ (x) ξ の標準偏差を表している。 Similarly, we have 同様にして、我々は以下の関係を得る、 ε(Q)σ(P) >= σ[N(Q)]σ[P(0)] > 1/2 |< [N(Q),D(P)] >| ・・・(30) σ(Q)η(P) >= σ[Q(0)]σ[D(P)] > 1/2 |< [Q(0),D(P)] >| ・・・(31) Thus, from Eqs. (28) - (31), we conclude Eq. (26). ここで、式(28) - 式(31) から、我々は式(26)という結論を得る。 --------------------------------------------------------------------------------