# How to Produce S-Tense Operators on Lattice Effect Algebras

## Abstract

Tense operators in effect algebras play a key role for the representation of the dynamics of formally described physical systems. For this, it is important to know how to construct them on a given effect algebra \( E\) and how to compute all possible pairs of tense operators on \( E\). However, we firstly need to derive a time frame which enables these constructions and computations. Hence, we usually apply a suitable set of states of the effect algebra \( E\) in question. To approximate physical reality in quantum mechanics, we use only the so-called Jauch–Piron states on \( E\) in our paper. To realize our constructions, we are restricted on lattice effect algebras only.

### Keywords

Effect algebra MV-algebra Complete lattice Tense operator S-tense operator Jauch–Piron E-state Jauch–Piron E-semi-state### Mathematics Subject Classification

Primary 03B44 03G25 06A11 06B23## 1 Introduction

Logic of quantum mechanics is an important tool for deciding and evaluation of propositions and propositional formulas on a physical system describing events in microcosmos. As known in quantum physics, behaviour of these physical systems and their elements (i.e., elementary particles) differs from physical systems which are observed in classical physics and whose behaviour is ruled by the classical propositional calculus. This was the reason that Foulis and Bennett [15] introduced the so-called effect algebras describing algebraic properties of propositions on events in quantum mechanics. For a more detailed motivation, the reader is referred to the monograph [12] by Dvurečenskij and Pulmannová.

Every physical system \(\mathcal P\) is a dynamic one which means that the true values of propositions about \(\mathcal P\) vary in time. However, the majority of propositional calculi used for a description of \(\mathcal P\) traditionally do not incorporate the time dimension. This means that the propositions are considered relative to a given moment. On the other hand, for a physical system \(\mathcal P\) there are usually discerned its states.

We assume that \(\mathcal P\) is in a state \(s_1\) at time \(t_1\) and it goes to a state \(s_2\) at time \(t_2\) (as an example may serve Minkowski spacetime \(\mathbb {R}^{1,3}\) where three ordinary dimensions of space are combined with a single dimension of time to form a four-dimensional manifold for representing a spacetime—elements of \(\mathbb {R}^{1,3}\) are then states in our setting). Hence this transition from the state \(s_1\) to \(s_2\) can be taken as a movement from time \(t_1\) to \(t_2\) in the time scale induced by the set \(S\) of states. Since not from every state \(s_1\) we can realize a transition to an arbitrary state of \(S\), there is a non-trivial binary relation \(R\) on \(S\) such that \((s_1,s_2) \in R\) if and only if the system \(\mathcal P\) can switch from \(s_1\) to \(s_2\). Hence, we take advantage to think of the set \(S\) of states as of the time scale and of the relation \(R\) as time preference, i.e., \((s_1,s_2) \in R\) if \(s_1\) is “before” \(s_2\) or \(s_2\) is “after” \(s_1\). Assuming that \((s_1,s_2) \in R\), \((s_2,s_3) \in R\) and possibly \(s_1=s_3\) means, among other things, that time loops are allowed. The couple \((S,R)\) is then the so-called time frame. Having a time frame of a physical system \(\mathcal P\), we can quantify our propositions in time as follows.

\(G(p)(s)\) is valid if for any \(t\) with \((s, t)\in R\) the formula \(p(t)\) is valid,

\(H(p)(s)\) is valid if for any \(t\) with \((t, s)\in R\) the formula \(p(t)\) is valid.

- (a)
if a time frame is given (or it is already constructed), how can we compute tense operators \(G^{*}\) and \(H^{*}\) on \( E\) having at least the same values for propositional formulas \(p(t)\) as by the given operators \(G\) and \(H\).

- (b)
how to construct all possible pairs of tense operators on \( E\).

- (c)
having tense operators satisfying our axiomatization on the lattice effect algebra \(E\) only, i.e., if a time frame is not given, under which conditions on \(E\) we can create the time frame in order to enable the solution of the above mentioned questions (a) and (b). This is usually called a

*representation problem*.

In fact, some solutions of above questions were already found by the authors, see e.g. [7, 16] and [17]. However, the next question is how our formal and purely algebraic approach coresponds to a real physical quantum system \(\mathcal P\).

For this sake, we prefer not to use all possible states of \(\mathcal P\) but only the so-called Jauch–Piron states. The advantage is that these states reflect an important property of the logic of quantum mechanics, namely the so-called Jauch–Piron property saying that if the probability of propositions \(p\) and \(q\) being true is zero then there exits a proposition \(r\) covering both \(p\) and \(q\) such that the probability of \(r\) being true is also zero. In our paper, we start with lattice effect algebras for the sake of simplicity. Hence, we can take \(r = p \vee q\). This is in accordance with Kolmogorovian probability theory. The study of Jauch–Piron states was motivated by the requirement that states on projection structures that qualify for having physical meaning should satisfy the natural Jauch–Piron property (see [18]).

- (a)
any time frame \((T, R) \) on a linearly ordered complete MV-algebra \(M \) induces S-tense operators on the cartesian product \(M^{T}\),

- (b)
conversely, any S-tense lattice effect algebra with a suitable set \(T\) of Jauch–Piron states is representable by some time frame \((T,R)\) as a subalgebra of the respective cartesian product \([0,1]^{T}\) equipped with the respective S-tense operators from (a).

- (c)
any frame \((S, T, R) \) on a linearly ordered complete MV-algebra \(M \) induces a very Jauch–Piron EM-function from \(M^{T}\) into \(M^{S}\).

## 2 S-Tense Operators on Lattice Effect Algebras

In this section we develop and study the basic concepts of tense effect algebras and an explanation of the role of time scale. We also give a justication for considering S-tense operators, describe their construction and a framework for their representation.

*effect algebra*is meant a structure \((E;+,0,1)\) where \(0\) and \(1\) are distinguished elements of \(E\), \(0\ne 1\), and \(+\) is a partial binary operation on \(E\) satisfying the following axioms for \(p,q,r\in E\):

- (E1)
if \(p+q\) is defined then \(q+p\) is defined and \(p+q=q+p\)

- (E2)
if \(q+r\) is defined and \(p+(q+r)\) is defined then \(p+q\) and \((p+q)+r\) are defined and \(p+(q+r)=(p+q)+r\)

- (E3)
for each \(p\in E\) there exists a unique \(p'\in E\) such that \(p+p'=1\); \(p'\) is called a

*supplement*of \(p\) - (E4)
if \(p+1\) is defined then \(p=0\).

*induced order*\(\le \) on \(E\) and the partial operation \(-\) as follows

It is worth noticing that \(a+b\) exists in an effect algebra \({E}\) if and only if \(a\le b'\) (or equivalently, \(b\le a'\)). This condition is usually expressed by the notation \(a\bot b\) (we say that \(a,b\) are *orthogonal*). Dually, we have a partial operation \(\cdot \) on \(E\) such that \(a\cdot b\) exists in an effect algebra \({E}\) if and only if \(a'\le b\) in which case \(a\cdot b=(a'+b')'\). This allows us to equip \(E\) with a dual effect algebraic operation such that \({E}^{op}=(E;\cdot ,1,0)\) is again an effect algebra, \('^{{E}^{op}}='^{{E}}='\) and \(\le _{{E}^{op}}=\le ^{op}\).

A *morphism of effect algebras* is a map between them such that it preserves the partial operation \(+\), the bottom and the top elements. A map \(s:E\rightarrow [0,1]\) is called a *state* on \({E}\) if \(s(0)=0\), \(s(1)=1\) and \(s(x+ y)=s(x)+s(y)\) whenever \(x+ y\) exists in \({E}\). Note that the real unit interval \([0,1]\) is an effect algebra such that \(x+ y\) exists in \([0,1]\) if and only if the usual sum of \(x\) and \(y\) is less or equal to \(1\).

*morphism*\(f:P_1\rightarrow P_2\)

*of bounded posets*is an order, top element and bottom element preserving map. Any morphism of effect algebras is a morphism of corresponding bounded posets. A morphism \(f:P_1\rightarrow P_2\) of bounded posets is

*order reflecting*if, for all \(a, b\in P_1\),

*isomorphism of effect algebras*is a surjective order reflecting morphism of effect algebras. In particular, \(':{{E}}\rightarrow {{E}^{op}}\) is an isomorphism of effect algebras.

If, moreover \((E;\le )\) is a lattice (with respect to the induced order), then \({E}\) is called a *lattice effect algebra*. On any lattice effect algebra \({E}\) we may introduce total operations \(\oplus \) and \(\odot \) as follows: \(x\oplus y=x+(y\wedge x')\) and \(x\odot y= (x'\oplus y')'\).

Given a positive integer \(n\in {\mathbb N}\), we let \(n\times x= x \oplus x \oplus x \cdots \oplus x\), \(n\) times, \(x^{n}= x \odot x \odot x \cdots \odot x\), \(n\) times, \(0x = 0\) and \(x^{0} = 1\).

Note that a lattice effect algebra \({E}\) is an *MV-algebra* (see [9]) with respect to the operations \(\oplus \) and \('\) if and only if \(x\wedge y=0\) implies \(x\le y'\). In this case are \(\oplus \) and \(\odot \) order preserving operations. It follows that any linearly ordered effect algebra is an MV-algebra.

*E-morphism between lattice effect algebras*is a morphism \(f:{{E}}_1\rightarrow {{E}}_2\) of effect algebras such that

- (i)
\(f(x\oplus x)=f(x)\oplus f(x)\).

- (ii)
\(f(x\odot x)=f(x)\odot f(x)\).

The axiomatization of tense operators \(G\) and \(H\) in the classical propositional logic is given in [3]. For effect algebras, it was settled in [6, 7] and [16]. We can repeat the definition.

*partial tense operators*if they are partial mappings of \(E\) into itself satisfying the following axioms:

- (T1)
\(G(0)=0\), \(G(1)=1\), \(H(0)=0\) and \(H(1)=1\),

- (T2)
\(x\le y\) implies \(G(x)\le G(y)\) whenever \(G(x), G(y)\) exist and \(H(x)\le H(y)\) whenever \(H(x), H(y)\) exist,

- (T3)
if \(x+y\) and \(G(x), G(y), G(x+y)\) exist then \(G(x)+G(y)\) exists and \(G(x)+G(y)\le G(x+y)\) and if \(x+y\) and \(H(x), H(y), H(x+y)\) exist then \(H(x)+H(y)\) exists and \(H(x)+H(y)\le H(x+y)\),

- (T4)
\(x\le GP(x)\) if \(H(x')\) exists, \(P(x)=H(x')'\) and \(GP(x)\) exists, \(x\le HF(x)\) if \(G(x')\) exists, \(F(x)=G(x')'\) and \(HF(x)\) exists.

*tense operators*and the triple \(({E};G,H)\) is called a

*tense effect algebra*.

This is exactly the case when a time frame \((T, R)\) is not explicitly mentioned. But it may happen, if a time scale \(T\) and a relation \(R\) for this axiomatization were constructed, that the obtained binary relation \(R\) is neither reflexive nor transitive. The conditions under which \(R\) will be a quasi-order are analysed in Theorems 1 and 3.

One can immediately mention that if \((E;+,0,1)\) is an effect algebra and \(G\) and \(H\) are mappings of \(E\) into itself defined by \(G(1)=1=H(1)\) and \(G(x)=0=H(x)\) for all \(x\in E\), \(x\not = 1\) then \(G\) and \(H\) are tense operators. However, these operators reveal little about the physical system \(\mathcal {P}\) because all that is not identically equal to \(1\) is considered to be false both in the past and in the future. Hence, we are searching to find tense operators on \({E}\) having maximally many values. The construction of an important class of such operators is given below in Theorem 15. However, it can happen that tense operators with maximally many values can loose their physical interpretation but another couple with smaller values can be more appropriate.

- (T5)
\(G(x\oplus x)=G(x)\oplus G(x), H(x\oplus x)=H(x)\oplus H(x)\),

- (T6)
\(G(x\odot x)=G(x)\odot G(x), H(x\odot x)=H(x)\odot H(x)\).

*E-tense operators*. Note that an

*embedding*between tense effect algebras is an order reflecting morphism of effect algebras such that it commutes with the corresponding tense operators.

Let us denote by \(S(E)=\{x\in E \mid x\wedge x'=0\}\) the set of all *sharp elements* of \(E\). Then, for any pair \(G\) and \(H\) of E-tense operators on \(E\) and any \(x\in S(E)\), we have that \(G(x)\) and \(H(x)\) are in \(S(E)\).

Now, in our case, by a *time frame* is meant a couple \((T,R)\) where \(T\) is a non-void set and \(R\subseteq T\times T\) is a reflexive and transitive relation. The last condition on \(R\) comes from the interpretion of \((s, t) \in R\) (or equivalently \(s R t\)) as \(s\) is before \(t\) in the non-strict sense so it would seem that \(R\) should be at least a quasi-order.

Having a lattice effect algebra \((E;+,0,1)\) and a non-void set \(T\), we can produce the direct power \((E^T;+, o,j)\) where the operation \(+\) and the induced operations \(\vee \), \(\wedge \), \(\oplus \), \(\odot \) and \(\lnot \) are defined and evaluated on \(x,y\in E^T\) componentwise. Moreover, \(o, j\) are such elements of \(E^T\) that \(o(t)=0\) and \(j(t)=1\) for all \(t\in T\). The direct power \({E}^T\) is again a lattice effect algebra.

The following theorem is an immediate corollary of Theorem 15 so we shall omit the proof up to Sect. 4.

**Theorem 1**

- (T7)
\(G(x)\le x, H(x)\le x\),

- (T8)
\(G(x)=G(G(x)), H(x)=H(H(x))\).

Motivated by Theorem 1 we say that E-tense operators \(G\) and \(H\) on a lattice effect algebra \(E\) are *S-tense* whenever they satisfy axioms (T7) and (T8) and that \(E\) is an *S-tense lattice effect algebra*.

It follows that \(G(x)\) means that “\(x\) is always going to be the case – starting now” and \(H(x)\) means that “\(x\) was always, and is now the case”. Moreover, the condition (T8) yields that the statement “\(x\) will always be the case” is equivalent to the statement “it will always be the case that \(x\) will always be the case”.

A *functional S-tense lattice effect algebra*\(E\) is an S-tense lattice effect algebra with S-tense operators \(G_E\) and \(H_E\) which is an E-effect-algebraic reduct of \([0,1]^{T}\) for a time frame \((T, R)\) such that \(G_E\) is a restriction of \(G\) from Theorem 1 on \(E\) and \(H_E\) is a restriction of \(H\) from Theorem 1, respectively.

By a *functional representation of an S-tense lattice effect algebra*\(A\) with S-tense operators \(G_A\) and \(H_A\), we mean simply a functional S-tense lattice effect algebra \(E\) such that there is an isomorphism of effect algebras \(f:A \rightarrow E\) satisfying \(f\circ G_A=G_E \circ f\) and \(f\circ H_A=H_E \circ f\).

As mentioned in Introduction, the set of states of an effect algebra can serve as a time scale over which we can quantify propositions on a physical system \({\mathcal P}\). Hence, it is worth to know as much as possible about states on a given effect algebra. For this, we introduce particular states, named E-states and Jauch–Piron E-states, which can be used for a representation of an effect algebra into a direct product of canonical effect algebras on the interval \([0,1]\). Although it can be introduced in full generality, we are concentrating only on the case of lattice effect algebras in order to reach everywhere defined mappings.

**Definition 2**

- (1)
*an E-state on*\({E}\) if \(s\) is an E-morphism of effect algebras; - (2)
*a Jauch–Piron E-state on*\({E}\) if \(s\) is an E-state and

- (JP)
\(s(x)=1= s(y)\) implies \(s(x\wedge y)=1\).

- (3)
If there exists an order reflecting set \(T\) of E-states (Jauch–Piron E-states, respectively) on \({E}\) then \({E}\) is said to be

*E-representable*(*E-Jauch–Piron representable*, respectively). - (4)
If any E-state is E-Jauch–Piron then \({E}\) is called an

*E-Jauch–Piron lattice effect algebra*.

First, note that any E-Jauch–Piron lattice effect algebra with an order reflecting set of E-states is E-Jauch–Piron representable. Also, any E-Jauch–Piron representable lattice effect algebra is E-representable.

Second, if \({E}\) is E-representable then the induced morphism \({i_{{E}}^T}:{{E}} \rightarrow [0,1]^{T}\) (sometimes called *an embedding*) is an order reflecting morphism of effect algebras such that \({i_{{E}}^T}(x\oplus x)={i_{{E}}^T}(x)\oplus {i_{{E}}^T}(x)\) and \({i_{{E}}^T}(x\odot x)={i_{{E}}^T}(x)\odot {i_{{E}}^T}(x)\) for all \(x\in E\).

Third, if there exists an order reflecting set \(T\) of states that are also lattice morphisms then \(E\) is an MV-algebra that is E-Jauch–Piron representable.

Fourth, if \({E}\) is an MV-algebra and \(s\) a state on \({E}\) then we always have \(s(x\vee y)+s(x\wedge y)=s(x)+s(y)\) for all \(x, y\in E\). Hence in every MV-algebra any state \(s\) satisfies (JP). Moreover from [2] we know that \(s\) is an E-state if and only if \(s\) is an extremal state (MV-algebra morphism).

The next theorem gives us a solution of the representation problem for S-tense operators. It immediately follows from Theorem 19 so we postpone its proof until Sect. 5.

**Theorem 3**

## 3 E-Semi-States on Lattice Effect Algebras

In this section we will systematically study the notion of an (Jauch–Piron) E-semi-state on a lattice effect algebra. The study of (Jauch–Piron) E-semi-states was motivated by the fact that a composition of an E-tense operator with an (Jauch–Piron) E-state is an (Jauch–Piron) E-semi-state. The basic result of this section is Proposition 11 that any Jauch–Piron E-semi-state on an E-representable lattice effect algebra \(E\) is a meet of E-states.

**Definition 4**

- (1)
*an E-semi-state on*\({E}\) if- (i)
\(s(0)=0, s(1)=1,\)

- (ii)
\(s(x)+ s(y)\le s (x+ y)\) whenever \(x+ y\) is defined,

- (iii)
\(s(x)\odot s(x)=s(x\odot x),\)

- (iv)
\(s(x)\oplus s(x)=s(x\oplus x),\)

- (i)
- (2)
*a Jauch–Piron E-semi-state on*\({E}\) if \(s\) is an E-semi-state and- (JP)
\(s(x)=1= s(y)\) implies \(s(x\wedge y)=1\).

- (JP)

**Lemma 5**

*Proof*

**Lemma 6**

- (a)
the pointwise meet \(t=\bigwedge S\) is an E-semi-state on \({E}\),

- (b)
if \(S\) is linearly ordered then \(q=\bigvee S\) is an E-semi-state on \({E}\).

*Proof*

(a): Let us check the conditions (i)–(iv) from Definition 4.

(i): Clearly, \(t(0)=\bigwedge \{s(0) \mid s\in S\} =\bigwedge \{0 \mid s\in S\}=0\) and \(t(1)=\bigwedge \{s(1) \mid s\in S\} =\bigwedge \{1 \mid s\in S\}=1\).

(i): Clearly, \(q(0)=\bigvee \{s(0) \mid s\in S\} =\bigvee \{0 \mid s\in S\}=0\) and \(q(1)=\bigvee \{s(1) \mid s\in S\} =\bigvee \{1 \mid s\in S\}=1\).

### 3.1 Dyadic Numbers and MV-Terms

We remark on some concepts introduced by Teheux in [23]. The set \(\mathbb D\) of dyadic numbers is the set of the rational numbers that can be written as a finite sum of integer powers of 2. We denote by \(f_0(x)\) and \(f_1(x)\) the terms \(x\oplus x\) and \(x\odot x\) respectively, and by \(T_{\mathbb D}\) the clone generated by \(f_0(x)\) and \(f_1(x)\).

*clone*is a set \(C\) of finitary operations on a set \(A\) such that

\(C\) contains all the projections \(\pi _{k}^{n}: A^n \rightarrow A\), defined by \(\pi _{k}^{n}(x_1, \ldots ,x_n) = x_k\),

\(C\) is closed under (finitary multiple) composition: if \(f, g_1, \ldots , g_m\) are members of \(C\) such that \(f\) is an \(m\)-ary operation, and \(g_j\) is an \(n\)-ary operation for every \(j\), then the \(n\)-ary operation \(h(x_1,\ldots ,x_n):= f(g_1(x_1, \ldots ,x_n), \ldots , g_m(x_1,\ldots ,x_n))\) is in \(C\).

**Lemma 7**

**Corollary 8**

**Proposition 9**

[2, Proposition 1] Let \({A}\) be a linearly ordered MV-algebra, \(s: A\rightarrow [0,1]\) an MV-morphism, \(x\in A\). Then \(s(x)=1\) iff \(t_r(x)=1\) for all \(r\in (0,1)\cap {\mathbb D}\).

Equivalently, \(s(x)<1\) iff there is a dyadic number \(r\in (0,1)\cap {\mathbb D}\) such that \(t_r(x)\not =1\). In this case, \(s(x)<r\).

Following the preceding corollary we are able to compare E-semi-states \(s\) via inclusion relation looking only on the respective order filters \(s^{-1}(\{1\}\).

**Lemma 10**

Let \((E;+,0,1)\) be a lattice effect algebra, \(s, t\) E-semi-states on \({E}\). Then \(t\le s\) iff \(t(x)=1\) implies \(s(x)=1\) for all \(x\in E\).

*Proof*

Clearly, \(t\le s\) yields the condition \(t(x)=1\) implies \(s(x)=1\) for all \(x\in E\).

Assume now that \(t(x)=1\) implies \(s(x)=1\) for all \(x\in E\) is valid and that there is \(y\in E\) such that \(s(y) < t(y)\). Thus, there is a dyadic number \(r\in (0,1)\cap \mathbb D\) such that \(s(y)<r< t(y)\). By Corollary 8 there is a term \(t_r\) in \(T_{\mathbb D}\) such that \(t_r(s(y))< 1\) and \(t_r(t(y))=1\). It follows that \(s(t_r(y))=t_r(s(y))< 1\) and \(t(t_r(y))=t_r(t(y))=1\). The last condition yields that \(s(t_r(y))=1\), a contradiction.

### 3.2 Riesz Decomposition Property and Ideals in Effect Algebras

We recall that an effect algebra \({E}\) satisfies the *Riesz Decomposition Property* ((RDP) in abbreviation) if \(x \le y_1 + y_2\) implies that there exist two elements \(x_1, x_2 \in E\) with \(x_1 \le y_1 \) and \(x_2 \le y_2\) such that \(x = x_1 + x_2\). Any MV-algebra satisfies (RDP).

*ideal*of an effect algebra \({E}\) is a non-empty subset \(I\) of \(E\) such that

- (i)
\(x \in E\), \(y \in I\), \(x\le y\) imply \(x \in I\),

- (ii)
if \(x,y \in I\) and \(x+y\) is defined in \({E}\), then \(x+y \in I\).

*filter*of an effect algebra \({E}\) is an ideal in the dual effect algebra \({E}^{op}=(E;\cdot ,1,0)\).

We denote by \({Id}({E})\) the set of all ideals of \({E}\). An ideal \(I\) is said to be a *Riesz ideal* if, for \(x \in I\), \(a,b \in E\) and \(x \le a+b\), there exist \(a_1,b_1 \in I\) such that \(x = a_1 +b_1\) and \(a_1 \le a\) and \(b_1 \le b\).

For example, if \({E}\) has (RDP), then any ideal of \({E}\) is Riesz and there is a one-to-one correspondence between congruences and ideals of \(E\) (see [13]), and given an ideal \(I\), the congruence \(\sim _I\) on \({E}\) is assigned by \(a\sim _I b\) iff there are \(x,y \in I\) with \(x\le a\) and \(y\le b\) such that \(a-x = b-y\). Then the quotient \({E}/I\) is an effect algebra with RDP.

The main supporting statement that we need (and which is essentially contained in the proof of [17, Theorem 4.5] or in [2, Proposition 3]) is

**Proposition 11**

Let \((E;+,0,1)\) be a lattice effect algebra with an order reflecting set \(S=\{ s: E\rightarrow [0,1]\mid \ s\ \text {is an}\ \text {E-state on}\)\({E}\}\), \(t\) a Jauch–Piron E-semi-state on \({E}\) and \(S_{t}=\{ s\in S \mid s\ge t\}\). Then \(t=\bigwedge S_t\).

*Proof*

We may assume that \(E\subseteq [0,1]^{S}\) such that \(x+y\), \(x\cdot y\), \(x\oplus x\) and \(x\odot x\) computed in \({E}\) gives us the same results as \(x+y\), \(x\cdot y\), \(x\oplus x\) and \(x\odot x\) computed in \([0, 1]^{S}\) and restricted to elements from \(E\). This means that the inclusion map \(i:{E}\rightarrow [0, 1]^{S}\) is an order reflecting E-morphism of lattice effect algebras. Note also that \([0, 1]^{S}\) is a lattice effect algebra with RDP (MV-algebra).

So we have that \(t_r(x)\notin V\). Let \(W\) be a maximal filter of \([0, 1]^{S}\) which does contain \(V\) and does not contain \(t_r(x)\). Then the set \(I=\{y\in [0, 1]^{S} \mid y'\in W\}\) is a prime ideal in \([0, 1]^{S}\), \(t_r(x)'\notin I\). It follows by [13, Proposition 6.5 and Proposition 6.10] that \([0, 1]^{S}/I\) is a linearly ordered effect algebra, i.e. an MV-algebra such that \(t_r([x]_{I})=[t_r(x)]_{I}\not =[1]_I\). In particular the factor map \(\pi _I:[0, 1]^{S}\rightarrow [0, 1]^{S}/I\) is an MV-morphism such that \(\pi _I(x)=[x]_{I}\) and \(\pi _I(U)\subseteq \pi _I(V)\subseteq \pi _I(W)=\{[1]_I\}\).

Let us denote by \(U_I\) the maximal ideal of \( [0, 1]^{S}/I\) and by \(\overline{s}: [0, 1]^{S}/I \rightarrow [0,1]\) the corresponding MV-morphism. Hence, by Proposition 9, \(\overline{s}([x]_{I})<r<1\). Let us put \(s=\overline{s}\circ \pi _I\circ i\). Then \(s\) is an E-state such that \(s(U)=\overline{s}(\pi _I(U))=\overline{s}(\{[1]_I\})=\{1\}\). It follows by Lemma 10 that \(t\le s\), i.e., \(s\in S_t\) and \(s(x)=\overline{s}([x]_{I})< r <\bigwedge S_t (x) \le s(x)\), a contradiction.

## 4 Functions Between Lattice Effect Algebras and Their Construction

This section studies the notion of an EM-function between lattice effect algebras and a very Jauch–Piron EM-function between lattice effect algebras. The overall goal of this section is to establish in some sense a canonical construction of very Jauch–Piron EM-functions. We would like to understand this construction because its particular case is the construction of S-tense operators when we apply the time frame \((T,R)\). This construction can serve as an ultimate source of numerous examples.

**Definition 12**

- (1)By an
*EM-function G between lattice effect algebras*is meant a function \(G:{E}_1\rightarrow {E}_2\) such that \((E_1;+_1, 0_1,1_1)\) and \(({E}_2;+_2, 0_2,1_2)\) are lattice effect algebras andIf \({E}_1={E}_2\) we say that \(G\) is an(EM1) \(G(0_1)=0_2\), \(G(1_1)=1_2\),

(EM2) \(G(x)+_2 G(y)\le G(x+_1 y)\),

(EM3) \(G(x)\odot _2 G(x)=G(x\odot _1 x)\),

(EM4) \(G(x)\oplus _2 G(x)= G(x\oplus _1 x)\).

*EM-operator*on \({E}_1\). - (2)If moreover \(G\) satisfies conditionswe say that \(G\) is an
(EM2)’ \(G(x)+_2 G(y)= G(x+_1 y)\)

*EM-morphism between lattice effect algebras*. - (3)If \(G\) is an EM-function between lattice effect algebras such that (EM5) (resp. (EM5)’) is satisfied we say that \(G\) is
*Jauch–Piron*(resp.*very Jauch–Piron*)(EM5) \(G(x)=1_2=G(y)\) implies \(G(x\wedge _1 y)=1_2\),

(EM5)’ \(G(x)\wedge _2 G(y)= G(x\wedge _1 y)\).

- (4)If \(G\) is an EM-function between lattice effect algebras such thatwe say that \(G\) is an
(EM6) \(G(x)\odot _2 G(y)\le G(x\odot _1 y)\),

(EM7) \(G(x)\oplus _2 G(y)\le G(x\oplus _1 y)\),

(EM8) \(G(x^{n})=G(x)^{n}\) for all \(n\in {\mathbb N}\),

(EM9) \(n\times _2 G(x)= G(n\times _1 x)\) for all \(n\in {\mathbb N}\),

*FEM-function*. - (5)
If \(G:{E}_1\rightarrow {E}_2\) and \(H:{B}_1\rightarrow {B}_2\) are EM-functions between lattice effect algebras, then a

*morphism between G and H*is a pair \((\varphi ,\psi )\) of morphisms of lattice algebras \(\varphi :{E}_1 \rightarrow {B}_1\) and \(\psi :{E}_2 \rightarrow {B}_2\) such that \(\psi (G(x))=H(\varphi (x))\), for any \(x\in E_1\).

The notion of an EM-function generalizes both the notions of an E-semi-state from Sect. 3 of this paper, of a \(\odot \)-operator from [17] for MV-algebras and of an fm-function from [2].

According to both (EM3) and (EM4), \(G|_{S({E}_1)}: {S({E}_1)} \rightarrow {S({E}_2)}\) is an EM-function (a Jauch–Piron EM-function, a very Jauch–Piron EM-function) whenever \(G\) has the corresponding property.

Recall also, that EM-functions between E-representable MV-algebras (which are exactly semisimple MV-algebras) coincide with FEM-functions between them (see [2, Proposition 6]). On the contrary, there is an MV-algebra \(M\) with an EM-operator on \(M\) that is not an FEM-operator (see [2, Proposition 7]).

**Lemma 13**

Let \(G:{E}_1\rightarrow {E}_2\) be an EM-function between lattice effect algebras, \(r\in (0,1)\cap {\mathbb D}\). Then \(t_r(G(x))= G(t_r(x))\) for all \(x\in E_1\).

*Proof*

Note that \(G(x)\oplus _2 G(x)= G(x\oplus _1 x)\) by (EM4) and \(G(x)\odot _2 G(x)= G(x\odot _1 x)\) by (EM3). Then, since \(t_r\in T_{\mathbb D}\) is defined inductively using only the operations \((-)\oplus (-)\) and \((-)\odot (-)\), we get \(t_r(G(x))=G(t_r(x))\).

By a *frame* is meant a triple \((S,T,R)\) where \(S,T\) are non-void sets and \(R\subseteq S\times T\). If \(S=T\), we will write briefly \((T, R)\) for the frame \((T,T,R)\). For our sake, we will assume that for every \(x\in S\) there is \(y\in T\) such that \(xRy\) and for every \(y\in T\) there is \(x\in S\) such that \(xRy\). Having a lattice algebra \((E;+,0,1)\) and a non-void set \(T\), we can produce the direct power \((E^T;+, o,j)\) where the operation \(+\) and the induced operations \(\vee \), \(\wedge \), \(\oplus \), \(\odot \) and \(\lnot \) are defined and evaluated on \(p,q\in E^T\) componentwise. Moreover, \(o, j\) are such elements of \(E^T\) that \(o(t)=0\) and \(j(t)=1\) for all \(t\in T\). The direct power \({E}^T\) is again a lattice effect algebra.

The notion of frame allows us to construct examples of FEM-functions between lattice effect algebras.

In what follows we will need the following

**Lemma 14**

- (1)\(\bigwedge \{a_i; i\in I\} + \bigwedge \{b_i; i\in I\}\) exists and$$\begin{aligned} \bigwedge \{a_i; i\in I\} + \bigwedge \{b_i; i\in I\}\le \bigwedge \{a_i+b_i; i\in I\} \end{aligned}$$
- (2)
\(\bigwedge \{c_i'; i\in I\}\le (\bigwedge \{c_i; i\in I\})'\).

We now prove the generalization of [2, Theorem 2].

**Theorem 15**

*Proof*

Trivially we can verify \(G(o)(s)=0, G(j)(s)=1\) for all \(s\in S\) due to the fact that \(o(t)=0\) and \(j(t)=1\) for each \(t\in T\) thus (EM1) holds. Let us check (EM2). Assume that \(p,q\in M^T\) and \(p+q\) exists. Hence, \(p(t)+q(t)\) exists for each \(t\in T\). Let \(s\in S\). By Lemma 14 (1) also \(\bigwedge \{p(t); s R t\} + \bigwedge \{q(t); s R t\}\) exists and \(G(p)(s)+G(q)(s)=\bigwedge \{p(t); s R t\} + \bigwedge \{q(t); s R t\}\le \bigwedge \{p(t)+q(t); s R t\}=G(p+q)(s)\). Thus \(G(p)+G(q)\le G(p+q)\).

The conditions (EM5)’, (EM6)-(EM9) and the adjointness between \(P\) and \(G\) follow directly from [2, Theorem 2], the condition (EM3) follows from (EM8) and the condition (EM4) follows from (EM9). The remaining part for \(P\) follows by the same arguments applied to the dual MV-algebra \({M}^{op}\).

We say that \(G:{M}^T \rightarrow {M}^S\) is the *canonical very Jauch–Piron FEM-function induced by the frame*\((S,T,R)\)*and the MV-algebra*\({M}\).

**Corollary 16**

Now we give the postponed proof of Theorem 1. It is enough to check conditions (T7) and (T8), the remaining parts follow immediately from Corollary 16. But (T7) and (T8) follow from [2, Theorem 6].

## 5 The Representation Theorem and its Applications

The aim of this section is to show that an E-representable E-Jauch–Piron lattice effect algebra with E-tense operators \(G\) and \(H\) (or with a very Jauch–Piron operator \(G\)) can be represented in a power of the standard MV-algebra \([0,1]\), where the set of all Jauch–Piron E-states serves as a time frame (in the sense given in Introduction) with a relation defined by means of the point-wise ordering of these states on \(x\) as well as on \(G(x)\) or \(H(x)\). It properly means that for every E-representable E-Jauch–Piron lattice effect algebra with tense operators a suitable time frame exists and can be constructed by use of the previously introduced concepts.

**Theorem 17**

Let \(G: E_1\rightarrow E_2\) be a very Jauch–Piron EM-function between lattice effect algebras. Let \( E_1\) be an E-representable lattice effect algebra and let \( E_2\) be an E-Jauch–Piron representable lattice effect algebra, \(T\) a set of all E-states from \( E_1\) to the standard MV-algebra \([0,1]\) and \(S\) a set of all Jauch–Piron E-states from \( E_2\) to \([0,1]\).

*Proof*

Assume that \(x\in E_1\) and \(s\in S\). Then \({i_{ E_2}^S}(G(x))(s)=s(G(x))\le t(x)\) for all \(t\in T\) such that \((s,t)\in \rho _G\). It follows that \({i_{ E_2}^S}(G(x))\le G^{*}({i_{ E_1}^T}(x))\).

Note also that [2, Theorem 3] is a particular case of Theorem 15.

The following result which is an immediate corollary of Theorem 17 generalizes the main result of the paper [17, Theorem 4.5].

**Corollary 18**

**(Representation theorem for lattice effect algebras with a very Jauch–Piron operator)** For any E-representable E-Jauch–Piron lattice effect algebra \( E\) with a very Jauch–Piron operator \(G\) and where T is an order reflecting set of all Jauch–Piron E-states, \( E\) is embeddable via E-morphism \(i_{ E}^{T}\) into the canonical MV-algebra \({ L}_G=([0,1]^{T};G^{*})\) with a very Jauch–Piron operator \(G^{*}\) induced by the canonical frame \((T,\rho _G)\) and the standard MV-algebra \([0,1]\). Further, for all \(x\in E\) and for all \(s\in T\), \(s(G(x))=G^{*}((t(x))_{t\in T})(s)\).

The next Theorem yields a solution of the representation problem for E-tense operators. The idea of the proof is taken from [2, Theorem 5].

**Theorem 19**

*Proof*

Due to the definition of \(\rho _H\) we have \(t\rho _H s\) and \(\rho _G\subseteq \rho _H^{-1}.\) Analogously we can prove the second inclusion.

The remaining part follows from Theorem 17. Basically, the obtained equations \(G^*(x) = G(x)\) and \(H^*(x)=H(x)\) finish the proof.

Now we give the postponed proof of Theorem 3. It is enough to show that the relation \(\rho _G\) from Theorem 19 is reflexive and transitive. But to prove this we can use the same arguments as in [2, Theorem 6] or [16, Lemma 4.7].

The next example shows an E-representable E-Jauch–Piron lattice effect algebra \(E\), that is not an MV-algebra, with all possible (E- and S-)tense operators. In particular, we constructed a (time) frame \((T,\rho )\) from the set \(T\) of all Jauch–Piron E-states on \(E\) and we also described a respective representation in \([0,1]^{T}\), in accordance to Theorem 19.

*Example 20*

Let us consider a lattice effect algebra \((E;+,0,1)\), where \(E = \{0, a, b, c, a + b = 2c = 1\}\) (see Fig. 1).

\(G_{1} = H_{1} = \mathrm{id}_{E}\),

\(G_{2}(a) = a, G_{2}(b) =0, G_{2}(c) = c \), \(H_{2}(a) = 0, H_{2}(b) =b, H_{2}(c) = c\),

\(G_{3}(a) = 0, G_{3}(b) =b, G_{3}(c) = c\), \(H_{3}(a) = a, H_{3}(b) =0, H_{3}(c) = c\),

\(G_{4}(a) = 0, G_{7}(b) =0, G_{7}(c) = c\), \(H_{7}(a) = 0, H_{7}(b) =0, H_{7}(c) = c\)

\(G_{5}(a) = b, G_{4}(b) =a, G_{4}(c) = c\), \(H_{4}(a) = b, H_{4}(b) =a, H_{4}(c) = c\),

\(G_{6}(a) = b, G_{5}(b) =0, G_{5}(c) = c\), \(H_{5}(a) = b, H_{5}(b) =0, H_{5}(c) = c\),

\(G_{7}(a) = 0, G_{6}(b) =a, G_{6}(c) = c\), \(H_{6}(a) = 0, H_{6}(b) =a, H_{6}(c) = c\),

There exist two extremal states (i.e., they cannot be written as a non trivial convex combination of other states) \(T = \{s_{1}, s_{2}\}\), given by \(s_1(a) = 1, s_{1}(b) = 0\), \(s_2(a)=0, s_2(b) =1\) and \(s_1(c) = s_2(c) = \frac{1}{2}\). The set \(T\) is also the set of all E-states, it is order reflecting and \(s_1, s_2\) are Jauch–Piron, hence \({ E}\) is an E-representable E-Jauch–Piron lattice effect algebra.

\(\rho _{G_{1}}:\)\(s_1 \rho _{G_{1}} s_1\), \(s_2 \rho _{G_{1}} s_2\),

\(\rho _{G_{2}}:\)\(s_1 \rho _{G_{2}} s_1\), \(s_2 \rho _{G_{2}} s_1\), \(s_2 \rho _{G_{2}} s_2\),

\(\rho _{G_{3}}:\)\(s_1 \rho _{G_{3}} s_1\), \(s_1 \rho _{G_{3}} s_2\), \(s_2 \rho _{G_{3}} s_2\),

\(\rho _{G_{4}}:\)\(s_1 \rho _{G_{7}} s_1\), \(s_2 \rho _{G_{4}} s_1\), \(s_2 \rho _{G_{7}} s_1\), \(s_2 \rho _{G_{7}} s_2\).

\(\rho _{G_{5}}:\)\(s_1 \rho _{G_{4}} s_2\), \(s_2 \rho _{G_{5}} s_1\),

\(\rho _{G_{6}}:\)\(s_1 \rho _{G_{5}} s_1\), \(s_1 \rho _{G_{6}} s_2\), \(s_2 \rho _{G_{5}} s_1\),

\(\rho _{G_{7}}:\)\(s_2 \rho _{G_{6}} s_2\), \(s_1 \rho _{G_{7}} s_2\), \(s_2 \rho _{G_{6}} s_1\),

\(G_{3}^{*}(i_{ E}^{T}(a)) = G_{3}^{*}(1,0) = (\bigwedge \{t(a) \mid t \in \{s_1,s_2\}, s_1 \rho _{3} t\}, \bigwedge \{t(a) \mid t \in \{s_1,s_2\}, s_2 \rho _{3} t\}) = (s_2(a),s_2(a)) = (0,0)\),

\(G_{3}^{*}(i_{ E}^{T}(b)) = G_{3}^{*}(0,1) = (\bigwedge \{t(b) \mid t \in \{s_1,s_2\}, s_1 \rho _{3} t\}, \bigwedge \{t(b) \mid t \in \{s_1,s_2\}, s_2 \rho _{3} t\}) = (s_1(b),s_2(b)) = (0,1)\),

\(G_{3}^{*}(i_{ E}^{T}(c)) = G_{3}^{*}(\frac{1}{2},\frac{1}{2}) = (\frac{1}{2},\frac{1}{2})\),

\(H_{3}^{*}(i_{ E}^{T}(a)) = H_{3}^{*}(1,0) = (\bigwedge \{t(a) \mid t \in \{s_1,s_2\}, t \rho _{3} s_1\}, \bigwedge \{t(a) \mid t \in \{s_1,s_2\}, t \rho _{3} s_2\}) = (s_1(a),s_2(a)) = (1,0)\),

\(H_{3}^{*}(i_{ E}^{T}(b)) = H_{3}^{*}(0,1) = (\bigwedge \{t(b) \mid t \in \{s_1,s_2\}, t \rho _{3} s_1\}, \bigwedge \{t(b) \mid t \in \{s_1,s_2\}, t \rho _{3} s_2\}) = (s_1(b),s_1(b)) = (0,0)\),

\(H_{3}^{*}(i_{ E}^{T}(c)) = H_{3}^{*}(\frac{1}{2},\frac{1}{2}) = (\frac{1}{2},\frac{1}{2})\).

## 6 Conclusion

Note that any semisimple MV-algebra is an E-representable E-Jauch–Piron lattice effect algebra. It follows that our result generalizes the results obtained in [2, 17]. Also, if we introduce modal operators on lattice effect algebras in accordance with [1] or [5] our method yields their representation on E-representable E-Jauch–Piron lattice effect algebras.

- (i)
a description of Jauch–Piron E-semi-states on lattice effect algebras with an order reflecting set \(T\) of E-states as meets of these E-states.

- (ii)
A representation theorem for E-representable E-Jauch–Piron lattice effect algebras with E-tense operators.

## Notes

### Acknowledgments

We thank the anonymous referees for the careful reading of the paper and the suggestions on improving its presentation. All authors acknowledge the support by ESF Project CZ.1.07/2.3.00/20.0051 Algebraic methods in Quantum Logic of the Masaryk University.

### References

- 1.Belluce, L.P., Grigolia, R., Lettieri, A.: Representations of monadic MV-algebras. Stud. Log.
**81**, 123–144 (2005)CrossRefMATHMathSciNetGoogle Scholar - 2.Botur, M., Paseka, J.: On tense MV-algebras. Fuzzy Set Syst (2014). doi:10.1016/j.fss.2014.06.006
- 3.Burges, J.: Basic tense logic. In: Gabbay, D.M., Günther, F. (eds.) Handbook of Philosophical Logic, vol. II, pp. 89–139. D. Reidel Publ. Comp. (1984)Google Scholar
- 4.Butnariu, D., Klement, E.P.: Triangular Norm-Based Measures and Games with Fuzzy Coalitions. Kluwer Academic Publishers, Dordrecht (1993)CrossRefMATHGoogle Scholar
- 5.Cattaneo, G., Ciucci, D., Dubois, D.: Algebraic models of deviant modal operators based on de Morgan and Kleene lattices. Inf. Sci.
**181**, 4075–4100 (2011)CrossRefMATHMathSciNetGoogle Scholar - 6.Chajda, I., Kolařík, M.: Dynamic effect algebras. Math. Slovaca
**62**, 379–388 (2012)CrossRefMATHMathSciNetGoogle Scholar - 7.Chajda, I., Paseka, J.: Dynamic effect algebras and their representations. Soft Comput.
**16**, 1733–1741 (2012)CrossRefMATHGoogle Scholar - 8.Chajda, I., Paseka, J.: Tense operators and dynamic De Morgan algebras, In: Proceedings of 2013 IEEE 43rd International Symposium Multiple-Valued Logic, pp. 219–224. Springer, Berlin (2013)Google Scholar
- 9.Cignoli, R.L.O., D’ottaviano, I.M.L., Mundici, D.: Algebraic Foundations of Many-Valued Reasoning. Kluwer, Dordrecht (2000)CrossRefMATHGoogle Scholar
- 10.Diaconescu, D., Georgescu, G.: Tense operators on MV-algebras and Łukasiewicz-Moisil algebras. Fundam. Inf.
**81**, 379–408 (2007)MATHMathSciNetGoogle Scholar - 11.Di Nola, A., Navara, M.: The \(\sigma \)-complete MV-algebras which have enough states. Colloq. Math.
**103**, 121–130 (2005)CrossRefMATHMathSciNetGoogle Scholar - 12.Dvurečenskij, A., Pulmannová, S.: New Trends in Quantum Structures. Kluwer Academic Publishers, Dordrecht (2000)CrossRefMATHGoogle Scholar
- 13.Dvurečenskij, A.: Perfect effect algebras are categorically equivalent with Abelian interpolation po-groups. J. Aust. Math. Soc.
**82**, 183–207 (2007)CrossRefMATHMathSciNetGoogle Scholar - 14.Figallo, A.V., Pelaitay, G.: Tense operators on De Morgan algebras. Logic J. IGPL (2013). doi:10.1093/jigpal/jzt024
- 15.Foulis, D.J., Bennett, M.K.: Effect algebras and unsharp quantum logics. Found. Phys.
**24**, 1325–1346 (1994)ADSCrossRefMathSciNetGoogle Scholar - 16.Paseka, J., Janda, J.: A dynamic effect algebras with dual operation. Math. Appl.
**1**, 79–89 (2012)MATHGoogle Scholar - 17.Paseka, J.: Operators on MV-algebras and their representations. Fuzzy Sets Syst.
**232**, 62–73 (2013). doi:10.1016/j.fss.2013.02.010 CrossRefMathSciNetGoogle Scholar - 18.PulmannovÁ, S., Pták, P.: Orthomodular Structures as Quantum Logics. Kluwer Academic Publishers, Dordrecht (1991)MATHGoogle Scholar
- 19.Riečanová, Z.: Generalization of blocks for D-Lattices and lattice-ordered effect algebras. Int. J. Theor. Phys.
**39**, 231–237 (2000)CrossRefMATHGoogle Scholar - 20.Riečanová, Z.: Basic decomposition of elements and Jauch–Piron effect algebras. Fuzzy Sets Syst.
**155**, 138–149 (2005)CrossRefMATHGoogle Scholar - 21.Rutledge, J.D.: A preliminary investigation of the in nitely many-valued predicate calculus. Ph.D. thesis, Cornell University (1959)Google Scholar
- 22.Teheux, B.: A Duality for the Algebras of a Łukasiewicz \(n+1\)-valued modal system. Stud. Log.
**87**, 13–36 (2007). doi: 10.1007/s11225-007-9074-5 CrossRefMATHMathSciNetGoogle Scholar - 23.Teheux B.: Algebraic approach to modal extensions of Łukasiewicz logics. Doctoral thesis, Université de Liege. http://orbi.ulg.ac.be/handle/2268/10887 (2009)