Arxiv:1901.01075V4 [Math.DS] 6 Jul 2021 Nrdcdi Ope Yaisb Eac 1]T Ieanwc New a Give to [10] Demarco by ( Dynamics Complex in Introduced Omlto Nscin5

Home , Arithmetic dynamics

arXiv:1901.01075v4 [math.DS] 6 Jul 2021 nrdcdi ope yaisb eac 1]t ieanwc new a give to [10] DeMarco by ( dynamics complex in introduced omlto nScin5. Section in formulation oecniuu pui)uhroi ucin eac sh DeMarco current) function. closed (pluri-)subharmonic (resp. continuous measure some positive locally-finite a is if h ako otlstermi i btcei h tde o ( studies of passivity the the in define to obstacle adopt big we Montel’s a n which of normality, statement In is weaker a theorem role. gave Montel’s essential [15] Favre-Kiwi-Trucco the of plays lack theorem the Montel’s this, show To oe analytically, moves euti ope yaisb an-u-ulvn[5 n i for and following: Mañe-Sud-Sullivan [25] the by states dynamics complex in result t rgnlydfie nti a;nml,fracmlxmanifol complex a for namely, way; this in defined originally betwee follo relation the the from moreover stability and the J-stability. them, consider between we sev relation paper, already the are this stabi there In the although theorem, of [22]. Montel’s definition of appropriate lack an the dynam giving non-archimedean in In even culties dynamics. motion of arithmetic stability in the problems i.e., J-stability, so-called the to frtoa ucin fdegree of functions rational of ups fti ae st eeo h tblt hoyfrs for theory stability the develop to is paper this of purpose aaerzdb mohstrictly smooth a by parametrized e scnie naayi aiyo ainlfunctions rational of family analytic an consider us Let 0 ,c f, { u eodapoc sb en fteso-called the of means by is approach second Our h rtapoc sb en fnraiy ncmlxdynami complex In normality. of means by is approach first The ncmlxdnmc,tesaiiy rthe or stability, the dynamics, complex In e od n phrases. and words Key 2010 Date Let ∈ f .Hr,teatvt esr,o h ciiycretfrah a for current activity the or measure, activity the Here, ). t n V ( K c 31/8/2019. : ahmtc ujc Classification. Subject Mathematics rjcieline projective Abstract. h adlrtst n otenraiyo h euneo th of stu sequence the and of normality functions, i the as rational to such of and locus activity set, family the Mandelbrot about a the more of study we point polynomials, critical a analytic by of parametrized polynomials of families of points ( ftefamily the if t ea lerial lsdfil ihcmlt,non-trivial complete, with field closed algebraically an be ) } n CIIYMAUE FDNMCLSSESOVER SYSTEMS DYNAMICAL OF MEASURES ACTIVITY snra around normal is oadteudrtnigo iucto hnmn fdynam of phenomena bifurcation of understanding the Toward f P { sJsal around J-stable is 1 ,an f o-rhmda yais iucto;Broihgeomet Berkovich bifurcation; dynamics; non-archimedean t n V ( vrnnacieenfils esuytesaiiy(rpass (or stability the study we fields, non-archimedean over c ( and t )) O-RHMDA FIELDS NON-ARCHIMEDEAN d } t f n 0 n akdpoint marked a and ntesneo 1] o eal,seDfiiin21.W trea We 2.16. Definition see details, For [15]. of sense the in saoe iha diinlasmto hteeycritical every that assumption additional an with above, as snra around normal is K rmr:3P0 eodr:1G2 63,37P30. 26E30, 14G22, Secondary: 37P50, Primary: aayi curve -analytic t 1. 0 EM IROKAWA REIMI fadol falteciia onsaepsiearound passive are points critical the all if only and if Introduction passivity 1 c t 0 : ,c f, and , V fJlast,adt h nieyintersection unlikely the to and sets, Julia of V ciiymeasure activity ftepi ( pair the of uvs ecntutteatvt measure activity the construct We curves. n akdpoint marked a and ;tepi ( pair the ); hc a ewitna h alca of Laplacian the as written be can which → f iyo ( of lity srlto obuddeslcs i.e., locus, boundedness to relation ts : hoe ymdfigtedfiiinof definition the modifying by theorem active c ar( pair a uch igtoapoce,adte study then and approaches, two wing P rlrslsb ivra 2]adLee and [29] Silverman by results eral wdta h upr fteactivity the of support the that owed yispoete.Frafml of family a For properties. its dy d owr orbit. forward e dpnetyb ybc 2] which [24], Lyubich by ndependently V 1 nacieendnmc,however, dynamics, on-archimedean h tblt norsneadthe and sense our in stability the n ( c,hwvr hr r eea diffi- several are there however, ics, V ge-iesoa aaee space, parameter igher-dimensional C × family a , tblt hoy saremedy, a As theory. stability f ,tepi ( pair the ), K niinfrpsiiyo h pair the of passivity for ondition n o-rhmda valuation. non-archimedean and fnt hr safundamental a is There not. if ,c f, ,c f, P s h asvt ftepi is pair the of passivity the cs, ,c f, 1 , ,wihmil oe from comes mainly which ), an swl-tde nrelation in well-studied is ) h ciiymaueis measure activity The . ,c f, spsiearound passive is ) ry. c nteBerkovich the on ics → f ). ,c f, P : vt)o critical of ivity) c 1 V , an : is ) × V fdegree of P passive → 1 ( C P ) 1 → , an around t The . d P 0 this t point 1 ∈ ≥ ( C t V 0 2 ) . 2 REIMI IROKAWA current coincides with the activity locus of (f, c). The importance of the activity measure comes not only from this fact but also from their nice potential-theoretic properties such as equidistribution as in [12]. In non-archimedean situation, on the other hand, we have good potentilal theory over smooth curves by Thuillier [30], whose discussion is almost parallel to that in complex case. By means of this, we will construct the non-archimedean activity measure µ(f,c) in Section 3; in case ft are polynomials, it can be written as the Laplacian of the function

1 n (1) h(f,c)(t) = lim log max 1, f (c(t)) on V. n→∞ dn | t | Our activity measure for (f, c) has the following properties:

Theorem 1.1 (Theorem 3.1, Proposition 3.7). Let K, V , f and c be as above. We set cn(t): = n 1 ∗ ft (c(t)). The sequence of positive measures dn cnδζ0,1 n converges weakly to the activity measure { } P1,an µ(f,c), where δζ0,1 is the Dirac mass at the Gauss point ζ0,1 in . Moreover, any exceptional point, i.e., a point in the set 1 := x P1,an c∗ (δ δ ) 0 weakly as n . E ∈ dn n x − ζ0,1 6→ →∞ is a type 1 point.

Having these two approaches, we next discuss the interrelations between them, and compare them with the notion of J-stability. To this end, we first introduce the boundedness locus for a certain algebraic family of monic polynomials with V = A1,an; when the characteristic of K is 0 or greater than d, we have a good parametrization of polynomials by critical points and the value at 0 (for details, see Section 4). In this case, we can consider the boundedness locus for the pair (f, c) with c critical: = t V c (t) is bounded . M(f,c) { ∈ |{ n } } Assuming is non-empty and bounded, we can show the following: M(f,c) Theorem 1.2 (Theorem 4.3). Let V , f, c and be as above (see Section 4 for the precise M(f,c) definition). Then, the activity measure µ(f,c) is the equilibrium measure of the set (f,c) with respect to . In particular, the support of the activity measure coincides with the boundaryM of . ∞ M(f,c) For the definition of the equilibrium measure, see Proposition 2.9. By using this property, we can compare the above two definitions of activity, and show a relation between passivity and J-stability by Silverman [29]. Proposition 1.3 (Proposition 5.2). Consider the same situation as in Theorem 1.2 and assume that its Mandelbrot set (f,c) is non-empty and bounded. Then the passivity locus coincides with the normality locus. M Corollary 1.4 (Proposition 5.3). Let K, V (= A1,an), and f be as in Proposition 1.3. Assume that the residue characteristic of K is 0 or greater than d. Then, every parameter t0 where f has an unstably indifferent periodic point lies in the activity locus of some critical point.

Silverman [29] showed that f is J-stable around t0 if and only if there exists a neighborhood U of t0 such that ft has no indiﬀerent periodic point or no type 1 repelling periodic point of multiplicity greater than 1 for all t. The above corollary, therefore, indicates a partial relation between the J-stability and the stability of (f, c). 1.1. Organization of this paper. In Section 2, we collect basic results on potential theory over Berkovich curves. In Section 3, we construct the activity measure and show an equidistribution result. In Section 4, we consider some family of polynomials and discuss a certain relation between the activity measure and the boundedness locus. In Section 5, we treat the passivity deﬁned by the normality and consider the relation between the two notions of activity, and moreover, the relation between them and the J-stability. We give an example in Section 6. ACTIVITY MEASURES OF DYNAMICAL SYSTEMS OVER NON-ARCHIMEDEAN FIELDS 3

1.2. Notations and conventions. Throughout the paper, K is an algebraically closed ﬁeld with a complete, non-trivial and non-archimedean valuation. We denote the ﬁbred product and ×K × the set-theoretic product. A smooth strictly K-analytic curve is a Berkovich analytic space which is smooth, good, paracompact, separated and purely of dimension 1.

Acknowledgements. I would like to thank my advisor Fumiharu Kato for constant support and helpful conversations. Also I would like to thank Professor Mattias Jonsson and Professor Yutaka Ishii for giving me advices on this paper and my researches. This research has been partially supported by RIKEN Junior Research Associate Program and JSPS Grant-in-Aid for JSPS Fellows, 20J14309.

2. Potential theory over smooth strictly K-analytic curves 2.1. Potential functions and finite Radon measures. For functions over a smooth and strictly K-analytic curve V , we can consider the Laplacian operator ddc. More precisely, the Laplacian operator is an R-linear map from the space of functions of bounded differential variation to that of finite Radon measures. To study the weak convergence of some given sequence of finite Radon measures, it is useful to consider the sequences of the potential functions corresponding to the Radon measures. Proposition 2.1 ([30] Théorème 3.3.13.). Assume that the variety V is proper and irreducible. Then, for any Radon probability measure µ and any point x V such that x is not classical and ∈ does not belong to the support Supp µ of µ, there exists the unique function ux,µ on V such that

ux,µ(x) = 0, and

ddc u = δ µ, x,µ x − where δx is the Dirac mass at the point x. Remark 2.2. For the deﬁnition of harmonic and subharmonic functions, see [30]. These are completely analogous to the ones in complex potential theory.

1 Example 2.3. Let z be an inhomogeneous coordinate of P . Let h∞(z) be a continuous function deﬁned by h (z) = log max(1, z ). ∞ | | This function is non-constant only on the path from the point to the Gauss point ζ0,1. From c ∞ this we have dd h∞(z) = δ∞ δζ0,1 and h∞(ζ0,1) = 0. The function h∞ is the potential function u . We denote this function− simply by u . − ζ0,1,δ∞ ζ0,1,∞ Similarly, we denote the potential function ux,δy for a non-classical x and any point y simply by ux,y. Later, we will regard them as two-variable function with respect to x and y. Example 2.4. In relation to the above functions, the following one is also important; for any 1,an 1,an 1,an ζ P , the generalized Hsia kernel δ(x,y)ζ : P P R with respect to ζ satisﬁes the following∈ equation: × → log δ(x,y) = u (y) u (ζ) u (ζ). ζ x,ζ0,1 − x,ζ0,1 − y,ζ0,1 Since we know that ddc u (y)= δ δ , we have x,ζ0,1 ζ0,1 − x ddc log δ(x,y) = δ δ , x ζ ζ − y c where ddx is the Laplacian operator with respect to the variable x, which means we take the

Laplacian as a function of x while y is considered to be constant. Since log δ(x,ζ)ζ = uζ0,1,ζ(ζ), the function log δ(x,y) u (ζ) is the potential function u (x) unless ζ is classical.− − ζ − ζ0,1,ζ ζ,y 4 REIMI IROKAWA

2.2. Pull-back operators. Let W be another strictly K-analytic curve and f : V W be a non-constant morphism. → ∗ Definition 2.5. In the above notation, let x be a non-classical point of W . the pull-back f δx of the Dirac mass at x by f is defined to be the locally finite measure ∗ f δx = mf (y)δy, −1 y∈fX{x} where mf (y) is the multiplicity of f at y. It is known in [30] that any subharmonic function h can be approximated by the decreasing c sequence hn n of smooth subharmonic functions, and the sequence of the Laplacians dd hn converges{ to the} Laplacian ddc h of h. From this fact, together with Theorem 2.1, we can{ extend} the pull-back operators to any finite positive Radon measures. Proposition 2.6 ([30], Proposition 3.2.13). Let V and W be strictly K-analytic curves, f : V W be a non-constant morphism, and g be a smooth and subharmonic function on W . Then, we→ have ddc(g f)= f ∗ ddc g. ◦ Note that, we can extend the same result to all the subharmonic function g, expressing it as a limit of smooth functions. Example 2.7 (canonical measure). For a polynomial φ of degree d, we can consider the canonical height function hφ associated to φ given by

1 n hφ(x) = lim log max(1, φ (x) ). n→∞ dn | | 1,an c This is a well defined, continuous and subharmonic function on P . Its Laplacian dd hφ is called the canonical measure, whose support is called the Julia set of φ. 2.3. Capacity theory for the projective line. Definition 2.8. On the Berkovich projective line P1,an, for any probability measure µ and a point ζ, the energy integral Iζ (µ) of µ with respect to ζ is defined to be

Iζ (µ)= log δ(x,y)ζ (x)dµ(x)dµ(y). P1,an P1,an − ZZ × For a compact set E P1,an ζ , let (E) be the set of probability measures supported on E. The ⊂ \ { } P Robin constant Vζ (E) of E with respect to the point ζ is deﬁned to be

Vζ (E)= inf Iζ (µ). µ∈P(E) The logarithmic capacity of E with respect to the point ζ is C(E; ζ)= e−Vζ (E). Proposition 2.9 ([1], Proposition 6.6 & Proposition 7.21). Assume that C(E; ζ) > 0. There exists an unique measure µ (E) such that E ∈ P Vζ(E)= Iζ(µE).

We call µE the equilibrium measure of E with respect to the point ζ. The potential function uζ,µE is also denoted by uζ,E. It might depend on the point ζ, but actually, we have the following fact: Proposition 2.10. If the compact set E is of capacity 0 with respect to some point ζ, i.e. C(E; ζ)= 0, then it is of capacity 0 with respect to any point ζ′. ACTIVITY MEASURES OF DYNAMICAL SYSTEMS OVER NON-ARCHIMEDEAN FIELDS 5

Hence, whenever we only care as to whether the capacity is zero, it is enough to discuss the capacity for some fixed point. We mainly consider the capacity with respect to in the later argument. ∞ The next proposition is used later: Proposition 2.11 ([1], Proposition 6.8 & Corollary 7.39). In the above notation, assume that E V ζ has positive capacity. Let U be a connected component of V E containing ζ. Then, ⊂ \ { } ζ \ the support of the equilibrium measure µE is contained in the boundary ∂Uζ of Uζ . Moreover, if uζ,E is continuous, the support of the measure µE coincides with ∂Uζ . 2.4. Arakelov-Green functions. On P1,an, we consider the following property to discuss Arakelov- Green functions: Definition 2.12. A probability measure µ is said to be of continuous potentials if for some non- classical point ζ the potential function uζ,µ is continuous. Note that it can be easily seen that if µ is of continuous potentials, then the potential function uζ,µ is continuous for any non-classical point ζ. Definition 2.13. For a probability measure µ with continuous potentials, an Arakelov-Green function g : P1,an P1,an R is defined to be µ × →

(2) gµ(x,y)= log δ(x,y)ζ dµ(ζ)+ C, P1,an − Z where C is some constant.

Of course we can deﬁne gµ(x,y) for any probability measure µ not necessarily with continuous potentials. The following properties, however, are important and valid only for ones with continuous potentials: Proposition 2.14 ([1], Proposition 8.66). For a probability measure µ with continuous potentials, the Arakelov-Green function gµ(x,y) is lower-semicontinuous as a function of 2-variables, continuous as a single-variable function of each variable, and symmetric with respect to x and y. It is of bounded diﬀerential variation, and ddc g (x,y)= δ µ. x µ y − Proposition 2.15 (Energy-minimizing principle, [1], Theorem 8.71). For a probability measure µ with continuous potentials, consider the energy integral

Iµ(ρ)= gµ(x,y)dρ(x)dρ(y), ZZV ×V for any probability measure ρ. Then, the integral takes its minimum if and only if ρ = µ. 2.5. Normality of functions over Berkovich curves. In complex dynamics, the following Mon- tel’s theorem plays a fundamental role; for a complex manifold V , any family of analytic functions from V to P1(C) which avoid 0, 1 and is normal. Unfortunately, it is known that this theorem is not true in non-archimedean setting.∞ In [15], however, they give an alternative deﬁnition of normality so that a similar result for the Montel’s theorem holds. Deﬁnition 2.16. Let X be an open subset of P1,an. A family of analytic functions on X with values in P1,an is normal if for any sequence f and anyF point x X, there exists a n ∈ F ∈ neighborhood V of x, and a subsequence fnj j which converges pointwise on V to a continuous function. { } We will use the next theorem, one of the consequences of the non-archimedean Montel’s theorem; 6 REIMI IROKAWA

Proposition 2.17 ([15], Theorem 2.1). Let U be a basic subset, i.e., the boundary of U consists of only type 2 or 3 points, and Y be an aﬃnoid domain of P1,an. Then any sequence of analytic functions fn : U Y admits a subsequence that converges pointwise on U to a continuous function. { → }

3. The construction of the activity measure Let V be a smooth strictly K-analytic curve, f : V P1,an P1,an be an analytic family of ×K → rational functions of degree d, and c : V P1 be a marked point. The goal of this section is as follows: →

Theorem 3.1. In the above settings, there exists a locally ﬁnite positive Radon measure µ(f,c), locally written as the Laplacian of some continuous subharmonic function, which satisﬁes the following property: For the family of analytic functions c (t) := f n(c(t)): V P1,an , we have { n t → }n 1 c∗ δ µ weakly as n . dn n 0,1 → (f,c) →∞ Moreover, if ft is a family of polynomials, the measure µ(f,c) can be globally written as the Laplacian of a continuous and subharmonic function on V .

Definition 3.2. The measure µ(f,c) defined above is called the activity measure of the pair (f, c). For a family of polynomials f and its critical point c = , Theorem 3.1 is easy; define the t 6 ∞ function h(f,c) by 1 h(f,c)(t) = lim − log max( cn(t) , 1). n→∞ dn | | Then, clearly the limit h(f,c) exists, and it is naturally extended to a continuous subharmonic function. Since c ∗ dd ( log max( cn(t) , 1)) = cn(δ0,1 δ∞) − | | ∗ − = cnδ0,1, we have the theorem for the families of polynomials. For general families of rational functions, we take a homogeneous lift of ft and define the potential function locally, and then patch their Laplacians together to get the required measure defined on whole V . For any point t V , take a sufficiently small affinoid neighbourhood ( ) of t in V such that ∈ M A qt(z) ft(z)= for some pt, qt [z]. pt(z) ∈A (1) (1) Let Pt(X,Y ) = Pt (X,Y ) and Qt = Qt (X,Y ) be a homogeneous lifts of pt and qt respectively (n) with z = Y/X, with non-zero resultant Res(Pt,Qt) = 0 for any t ( ). Define Pt = (n−1) (1) (n) (n−1) 6 (1) ∈ M A Pt (Pt,Qt) and Pt = Pt, and Qt = Qt (Pt,Qt) and Qt = Qt respectively. Note that we (n) (n) have Res(Pt ,Qt ) = 0 for any t. Next, let us consider6 a lift C(t): V A2,an 0 of c(t): V P1,an with no pole as follows; for a sufficiently small neighborhood →, we may\ { assume} c(t) →A1,an for any t by changing the coordinate if necessary. The lift is definedA by C(t) = (1, c(t)). ∈ Now we construct locally the activity measure. Take any t V , an affinoid neighbourhood ( ) of t, and an open neighborhood U of t in ( ). Set ∈ M A M A 1 h(n)(t)= log P (n)(C(t)),Q(n)(C(t)) , n t t where , is a generalized Hsia kernel with respect to ζ0,1; note that it is an upper-semicontinuous functionk· of·k two variables satisfying x,y = max( x , y ) for any x,y K. k k | | | | ∈ ACTIVITY MEASURES OF DYNAMICAL SYSTEMS OVER NON-ARCHIMEDEAN FIELDS 7

Proposition 3.3. In the above notation, for any t U(K), we have positive functions C (t) and ∈ 1 C2(t) such that C (t) x,y d P (x,y),Q (x,y) C (t) x,y d, 1 k k ≤ k t t k≤ 2 k k for all t U of type 1 and all (x,y) A2(K). ∈ ∈ Proof. Write P (z)= a (t)zi and Q (z)= b (t)zi. Then we have C (t) = max ( a (t) , b (t) ). t i i t i i 2 i,j | i | | j | For C1(t) we need a little more complicated argument. First of all, since Res(Pt,Qt) = 0 there P P 6 exist polynomials G1,t, G2,t, H1,t, and H2,t Z[ ai(t) , bi(t) ,X,Y ] homogeneous in X and Y of degree d 1 such that ∈ { } { } − 2d−1 PtG1,t + QtG2,t = Res(Pt,Qt)X , and 2d−1 PtH1,t + QtH2,t = Res(Pt,Qt)Y . Taking the maximum of the absolute values of both sides of these equations, we have max( P G , Q G , PH , QH ) Res(P ,Q ) max( X , Y ). | t 1,t| | t 2,t| | 1,t| | 2,t| ≥ | t t | | | | | Since G1,t, G2,t, H1,t, and H2,t Z[ ai(t) , bi(t) ,X,Y ] are homogeneous of degree d 1, we have ∈ { } { } − G (X,Y ) max( a (t) , b (t) ) max( X , Y )d−1, | 1,t |≤ | i | | i | · | | | | and the similar inequality holds for G2,t, H1,t and H2,t. Therefore, max( P , Q ) max( a (t) , b (t) ) max( X , Y )d−1 Res(P ,Q ) max( X , Y )2d−1. | t| | t| | i | | i | | | | | ≥ | t t | | | | | For each t, taking B(t) = max( a (t) , b (t) ) > 0, we have | i | | i | B(t) P (X,Y ),Q (X,Y ) Res(P ,Q ) X,Y . k t t k ≥ | t t | · k k Therefore, by setting C (t)= Res(P ,Q ) /B(t) we have 1 | t t | C (t) x,y d P (x,y),Q (x,y) . 1 k k ≤ k t t k

Proposition 3.4. In the above notation, the sequence h(n) converges uniformly to a continuous { } and subharmonic function (denoted by h(f,c)). Proof. Taking (x,y)= C(t) we have (3) C (t) C(t) d P (C(t)),Q (C(t)) C (t) C(t) d 1 k k ≤ k t t k≤ 1 k k for all t U of type 1. Since C and C are determined by taking maximum, multiplication, addi- ∈ 1 2 tion, and division of coefficients of Pt and Qt. the functions log C1(t) and log C2(t) can be extended to whole U continuously. Then, since the set of type 1 points is dense in U, we have the formula (3) for all t U. Therefore, for each compact subset E of U, letting C = log maxt(C1(t),C2(t)), we have ∈ 1 log C(t) log P (C(t)),Q (C(t)) + C, k k≤ d k t t k which shows the locally uniform convergence of h(n) on E. { } Continuity and subharmonicity follow from those of h(n) { } Define the measure µU = ddc h on U. This is a positive finite Radon measure on U. (f,c) − (f,c) n Proposition 3.5. In the above notations, define cn(t)= ft (c(t)). Then, we have

1 ∗ (cn U ) δζU µ(f,c) as n on U, dn | 0,1 → →∞ where the limit is weak convergence. 8 REIMI IROKAWA

Proof. we have 1 h(n)(t) = log max( P (n)(C(t)) , Q(n)(C(t)) ) dn | t | | t | 1 (4) = log max( f n(c ) , 1) + log max( P (n)(C(t)) , 1) + log c (t) , dn | t t | | t | | 1 | where C(t) = (c1(t), c2(t)) for all t V t : ft(c(t)) = or c(t) = . We can extend the equality if we take the limit for each∈ point\ we { exclude. As we∞ see in Example∞} 2.3 we have ddc log P (n)(C(t)) = c∗ (ddc δ δ ) on U. t n ∞ − ζ0,1

For the third term of (4), we have

ddc log c (t) = a(x)δ , | 1 | x x∈XV (K) where div c1 = a(x)[x]. We have the similar formula for the second one. By the hypothesis on the lifts of ft and ct, we know that P c2 has no pole, • if c has a zero, then P (c(t)) has a pole of the same degree • 2 | t | if Pt(C(t)) has a zero, then ft(c(t)) has a pole of the same degree, • P|(C(t)) and| Q (C(t)) has| no pole.| • | t | | t | Therefore, we have 1 1 ddc log max( f (c ) , 1) + log max( c(t) , 1) + log c (t) = c∗ δ , dn | t t | | | | 1 | −dn n ζ0,1 where c (t)= f n(c(t)). As h(n) h on U locally uniformy as n , we have n t → (f,c) →∞ 1 ∗ U (5) lim c δζ = µ weakly, n→∞ dn n 0,1 (f,c) which is a positive ﬁnite Radon measure. U Proposition 3.6. The above measure µ(f,c) is independent of the choice of the open set U and ′ ′ ′ ′ the lifts Pt, Qt and C(t), i.e., if U is another open set and Pt , Qt and C (t) is a lift of (f, c) on ′ U satisfying the same condition as Pt, Qt and C(t), then U U ′ µ ′ = µ ′ . (f,c)|U∩U (f,c)|U∩U Proof. Since both of C(t) and C′(t) have no zero or pole on U and U ′ respectively, there exists an invertible analytic function ϕ on U U ′ such that ∩ C′(t)= ϕ(t) C(t) on U U ′. · ∩ ′ ′ For (Pt,Qt) and (Pt ,Qt), since both of them represents the same rational function, we have ′ Pt (X,Y )= a(t)Pt(X,Y ), and Q′ (X,Y )= a(t)Q (X,Y ) on U U ′, t t ∩ where a(t) is an invertible analytic function on U U ′. Note that since ϕ and a(t) are invertible, log ϕ and log a are harmonic. Therefore, we have∩ | | | | ′(n) ′ ′(n) ′ log max Pt (C (t)) , Qt (C (t)) (n) ( n) = log max P (C(t)) , Q (C(t)) + d n log ϕ(t) + (dn−1 + 1) log a(t) . t t | | | | ′ n for any n = 1, 2, 3,... on U U . Dividing them by d and taking a limit, we have ∩ h (t)= h′ (t) + log ϕ(t) + log a(t) on U U ′, (f,c) (f,c) | | | | ∩ ACTIVITY MEASURES OF DYNAMICAL SYSTEMS OVER NON-ARCHIMEDEAN FIELDS 9

′ ′ ′ ′ where h(f,c)(t) is defined in the same way as h(f,c) replacing Pt, Qt, and C by Pt , Qt and C respectively. As a consequence, we have ddc h = ddc h′ on U U ′. (f,c) (f,c) ∩ The above proposition shows the compatibility of the system of locally defined measures µU . { (f,c)} Hence we can glue them together to get the global measure µ(f,c) defined on the whole V . Since the formula (5) is valid locally, this holds globally on V . This completes the proof of Theorem 3.1. Note that the activity measure is defined independently of the choice of the Gauss point. More generally, we have the following: Proposition 3.7. For any non-classical point ζ P1,an, we have ∈ 1 c∗ δ µ as n , dn n ζ → (f,c) →∞ where the limit is weak convergence.

Proof. For any point ζ, deﬁne the function uζ to be

uζ(x) = log δ(x,ζ)ζ0,1 deﬁned in Example 2.4. This is a bounded and continuous function on P1,an, whose Laplacian is c R dd uζ = δζ0,1 δζ . For each n, consider uζ (cn(t)) : V . This is also a bounded continuous function satisfying− ddc u (c (t)) = c∗ (δ δ ). Since → ζ n n ζ0,1 − ζ 1 lim uζ (cn(t)) = 0 n→∞ dn 1 ∗ uniformly on V , we have n c (δ δ ) 0 weakly as n . d n ζ0,1 − ζ → →∞ Remark 3.8. The exceptional set 1 := x P1,an (δ δ ) 0 as n E ∈ | d x − ζ0,1 6→ →∞ is actually of capacity 0. We show this fact in a forthcoming paper with Y. Okuyama.

4. The activity measure vs. the boundedness locus In this section, we will consider a relation between the activity measure and the boundedness locus for a certain family of polynomials. 4.1. The boundedness locus. In this section, we consider the following situation. We assume that the characteristic of K is 0 or greater than d. For c = (c ,...,c ) Kd−2 and a K, define 1 d−2 ∈ ∈ a polynomial fc,a of degree d as follows: d−2 1 d σk(c1,...,cd−2) d−k d fc (z)= z + z + a , ,a d d k Xk=1 − where σk is the k-th elementary symmetric polynomial. Then, the critical points of fc,a are c0 = d−2 0, c ,...,c and fc (0) = ad. Regarding the set A A of points (c, a) as a parameter space 1 d−2 ,a K ×K K of polynomials, we consider its compactification P := Pd−1 and its boundary P := P (Ad−2 K ∞ \ K ×K A ). The family we consider in this section is a 1-parameter subfamily of this Ad−2 A given K K ×K K as follows; take any line V of P such that V P∞ consists of a single point denoted by . Let V denote the subset of the parameter space Ad∩−2 A defined by V = V , which we∞ identify ×K \ {∞} with A1 by an affine coordinate t. By abuse of notation, we denote by V and V their analytification respectively. Let ft = fc(t),a(t) denote the analytification of the restriction to V of the map fc,a. For a marked point, take one of ci’s and denote by c its analytifictaion. We consider the relation between the activity and the boundedness of the orbit of c by f i.e. c (t) . { n }n 10 REIMI IROKAWA

We consider the boundedness locus of a given critical point, which is a higher-degree analogue of Mandelbrot set: for any pair (f, c) considered above, deﬁne = t V f n(c(t)) ∞ is bounded . Mf,c { ∈ |{ t }n=0 } 2 1,an Remark 4.1. Consider the family of quadratic polynomials ft(z) = z + t (t A ). In this case, the point 0 is the unique critical point for any t and we have the boundedness∈ locus, i.e. the “Mandelbrot set” = t A1,an f n(0) ∞ is bounded . M { ∈ |{ t }n=0 } This set is, in non-arhimedean case, actually, just the closed unit disc (for details, see Section 6). The above f,c is, hence, despite with slightly diﬀerent parametrization, can be seen as some sort of higher degreeM generalization of the Mandelbrot set.

From now on, we assume that the set f,c is a non-empty bounded set of the parameter space V . This condition is not extremely strong,M as we discuss in Section 4.2. In this case we have the following:

Proposition 4.2. If f,c is non-empty, the activity measure µ(f,c) is a strictly positive ﬁnite Radon measure. M

Proof. The function h(f,c) is defined as a non-constant function on V since h(f,c)(t) = 0 for any t on 1,an the boundedness locus f,c and h(f,c) > 0 for any t on the non-boundedness locus P f,c by definition and since we assumeM that the both sets are non-empty. We have thus the activity\ Mmeasure is a strictly positive finite Radon measure. Indeed, If the activity measure µ(f,c) is zero measure, 1,an the function h(f,c) should be constant since every harmonic function on A is constant. The relation between this set and the activity measure is the following:

Theorem 4.3. In the above notation, we assume f,c is non-empty and bounded. Then, the activity measure µ coincides with the equilibrium measureM µ of the set with respect to (f,c) Mf,c Mf,c times µ(f,c)(V ). Moreover, the support of the activity measure coincides with the boundary of ∞ . Mf,c Remark 4.4. For the definition of equilibrium measure, see Section 2. This is well-defined only when • the set f,c is compact and of positive capacity. Since the positivity of the capacity is shown later,M let us see the compactness of the set here. Since is the zero locus of the Mf,c dynamical Green function h(f,c) and it is continuous, it is closed. Also we assumed that the set is bounded. Hence it is also closed in the projective line P1,an, which is compact and Hausdorff. As µ is positive and finite, we get a probability measure by taking a normalization • (f,c) µ(f,c)/µ(f,c)(V ). The above theorem states this probability measure coincides with the equilibrium measure of . Mf,c To prove the theorem, we need several lemmas. Before stating them, let us prepare some no- 1,an tations. Since the function cn is algebraic, it can be extended to a function V P where V = V as we define in the beginning of this section. We can also extend the→ function h ∪ {∞} (f,c) to a function defined on V by

h(f,c)( )= limsup h(f,c)(x). ∞ x→∞,x∈V

We denote these extensions by cn and h(f,c), respectively, by abuse of notation. For the activity measure µ = µ(f,c), there exists a probability measureµ ˆ deﬁned on V such that µ(f,c) µˆ V = . | µ(f,c)(V ) ACTIVITY MEASURES OF DYNAMICAL SYSTEMS OVER NON-ARCHIMEDEAN FIELDS 11

Indeed, since the extended dynamical Green function h(f,c) is of bounded diﬀerencial variation, we can consider the Laplacian of h(f,c), which gives us the formula ddc h = δ µˆ on V . (f,c) ∞ − Lemma 4.5. The support of µˆ is contained in the boundary of the set . Mf,c P1,an ◦ Proof. Note that Supp(ˆµ) f,c since h(f,c) 0 on f,c. Hence it is enough to show 1⊂,an \ M ≡ M that for every point t0 P f,c, there exists an open neighborhood U f,c of t such that ∈ d \ M i ⊂ M µˆ U is zero. Set ft(z) = i=0 ai(t)z with ad(t) = 0 and take a positive number M such that | 6 P1,an M > max0≤i≤d−1 ai(t0) , 1 . Then, for every z with z > M, the canonical height of ft0 coincides with the{| naive one,| P} i.e. ∈ | |

(6) hf (z)= z . t0 | | Indeed, for these z, we have z d > a (t)zi for any i = 0,...d 1, so | | | i | − h (z) = log max 1, a (t)zi ft0 i = log zd = d logXz | | from the ultrametric inequality. Take an open neighborhood U of t0 so small that M > max0≤i≤d ai(t) {| |} for any t U. Then the equation (6) holds for all t U if we replace t0 by t. Since the set f n(c(t)) ∈is unbounded by the deﬁnition of , there∈ exists a natural number N such that { t } Mf,c 0 f N0 (c(t )) > M. Taking a smaller open neighborhood of t in U if necessary, we can assume that | t0 0 | 0 for ever t U we have f N0 (c(t)) > M. Hence together with (6) we have ∈ | t | 1 N0 h(f,c)(t)= log max(1, ft (c(t)) ). dN0 | | Since c 1 ∗ dd h(f,c) = deg(cN0 )δ∞ cN δζ0,1 dN0 − 0 A1 A1 on U where deg cN0 is a degree of cN0 : V as a polynomial, and we have h(f,c)(t0) > −1 ≃ → c 0= h(f, c)(t) for any t cn ζ0,1 . Hence we can take smaller U to get dd h(f,c) =0 on U. ∈ { }

Therefore, we haveµ ˆ = µ(f,c) on V . Also, we have the following by the above lemma: Lemma 4.6. The measure µ has continuous potentials, i.e., for any non-classical point ζ, the potential function uζ,µ is continuous. Proof. Note that the measure µ can be written as the Laplacian of h on V and as the Laplacian − (f,c) of u(x)=0on V f,c, both of which are continuous subharmonic function. The claim follows from [1, Proposition8.65]\ M

Now we prove Theorem 4.3. Let us consider the Arakelov-Green function gµˆ(x,y) discussed in Section 2.4. This can be written explicitly by means of the dynamical Green function: Lemma 4.7. There exists a real constant C such that (7) g (x,y)= h (x)+ h (y) log δ(x,y) + C. µˆ (f,c) (f,c) − ∞ Proof. Take the Laplacian of each side with respect to x: ddc g (x,y) = δ µ,ˆ and x µˆ y − ddc (h (x)+ h (y) log δ(x,y) = δ (ˆµ δ δ ) x (f,c) (f,c) − ∞ ∞ − − ∞ − y = δ µ.ˆ y − 12 REIMI IROKAWA

Hence there exists a constant C(y) depending on y such that g (x,y)= h (x)+ h (y) log δ(x,y) + C(y). µˆ (f,c) (f,c) − ∞ Take the Laplacian with respect to y, we have that C(y) is constant, which shows the lemma.

The former part of Theorem 4.3 follows from the next lemma;

Lemma 4.8. In the above settings, the two measures µˆ and µMf,c coincide.

Proof. We denote by Iµˆ(ρ) the energy integral for a probability measure ρ:

Iµˆ(ρ)= gµˆ(x,y)dρ(x)dρ(y), P1,an P1,an ZZ × which takes its minimum if and only if ρ =µ ˆ by proposition 2.15. The deﬁnition of the equilibrium measure is the minimizer of another energy integral I∞; as in Section 2.3, for a probability measure ρ supported on , Mf,c I (ρ)= u (x)dµ(x)dµ(y) ∞ − ∞,y ZZMf,c×Mf,c takes its minimum if and only if ρ = µMf,c . Now we compare the above two energy minimizing principle and Lemma 4.7. Since the support ofµ ˆ is contained in ∂ f,c by lemma 4.5, we haveµ ˆ ( f,c). Since h(x) = 0 on the set , we have by (7) M ∈ P M Mf,c

Iµˆ(ρ) = gµˆ(x,y)dρ(x)ρ(y) ZZ = (h(x)+ h(y) + log δ(x,y)∞ + C)dρ(x)ρ(y) ZZ (8) = log δ(x,y)∞dρ(x)ρ(y)+ C = I∞(ρ)+ C ZZ for any ρ ( ). sinceµ ˆ ( ), we see that V (E) I (ρ) C < which shows the ∈ P Mf,c ∈ P Mf,c ∞ ≤ µˆ − ∞ well-deﬁnedness of the measure µMf,c . Moreover, Iµˆ(ρ) takes its minimum if and only if ρ =µ ˆ, while I∞(ρ) does if and only if ρ = µMf,c . Therefore we getµ ˆ = µMf,c .

Sinceµ ˆ has continuous potentials, so is µMf,c , which implies Supp µMf,c = ∂ f,c by proposition 2.11. This concludes the proof of Theorem 4.3. M

4.2. Boundedness condition for the boundedness locus (f,c). In this subsection, we give a suﬃcient condition for the boundedness locus to be bounded.M M(f,c) Proposition 4.9. For j = 0, 1,...,d 2, deﬁne = . Then, − Mj Mf(c,a),cj

P H := [c : a : 0] f c (c (c, a)) = 0 , Mj ∩ ∞ ⊂ j ( ,a) j

Pd−2,an where E is the topological closure of E in K for a set E.

Admitting the above proposition, we have a suﬃcient condition for j to be bounded; if j does not intersect with H , the set is bounded. Since H is a hypersurfaceM of P , for a generalM j Mj j ∞ choice of the line V of P, the boundedness locus j is bounded. To show the proposition, we need several lemmas;M

Lemma 4.10. f(c,a)(cj) = 0 for any j if and only if (c, a) = (0,..., 0, 0). ACTIVITY MEASURES OF DYNAMICAL SYSTEMS OVER NON-ARCHIMEDEAN FIELDS 13

Proof. First, as f(c,a)(0) = 0, we have a = 0. Note that for any ci with f(c,a)(ci) = 0, the multiplicity m(ci) is 1 + (number of j s.t. ci = cj) since the set of the critical points of f(c,a)(z) is c0(= 0), c1,...,cd−2 and the multiplicity is the order of f(c,a)(0) = 0 at 0. Suppose there exists a{ non-zero critical point.} Then, d = m(z)= m(c ) 2 + (d 1), j ≥ − f(c,aX)(z)=0 X(∗) where (*) is the set of the distinct critical points. As this is contradiction, we have c = (0,..., 0). Note that this is a purely algebraic result, so we have the same result for an arbitrary field. For (c, a) Kd−1, define following functions: ∈ 1 + g(c,a)(z) := lim log f(c,a)n (z) , n dn | | gj(c, a) := g(c,a)(cj), and G(c, a) := max g (c, a) , { j } where log+ z = log max(1, z ) for any z K. | | | | ∈ Lemma 4.11. G(c, a) = max c , a , 1 . {| j | | | } αj α0 Proof. Set cj = π uj for j = 1,...d 2 and a = π u0, where αj, α0 is a real number, π is an · − · + element in K such that π < 1, and uj = 1. As the claim is trivial when log max cj , a = 1, + | | | | {| | | |} we assume that log max cj , a > 1. Let I 0,...,d 1 be the set of all the indices that αj {| | | |} ⊂ { − } π attains the max in G(c, a). Then, G(c, a) is attained by some of cj with j I. Assume first 0| I| . For any j I, we have ∈ 6∈ ∈ d−2 d−i 1 d ( 1) σi(c) d−i d f c (c ) = c + − c + a ( ,a) j d j d i j i=1 X − d−2 1 ( 1)d−iσ ( u , 0,..., 0) dαj d i k k∈I d−i = π uj + − { } uj , · d d i i=1 − X where σi( u , 0,..., 0) is the i-th fundamental symmetric polynomial of degree d generated by { k}k∈I uk k∈I and d 2 I times of 0’s (considered as indeterminates). By Lemma 4.10, there exists j{ }I such that− − | | ∈

d−2 1 ( 1)d−iσ ( u˜ , 0,..., 0) (9) u˜d + − i { k}k∈I u˜d−i = 0, d j d i j 6 Xi=1 − whereu ˜k is the reduction of uk modulo the maximal ideal of K. For this j, we have f(c,a)(cj) = d d | | c . Now, since it is easy to see that f c (z) = z for any z with z > max c , a , We have | j| | ( ,a) | | | | | {| i| | |} 1 G(c, a)= g (c, a)= log f c (c ) = log c = max c , a . j dn ( ,a) j | j | {| i| | |} when j = 0 is the non-zero solution of (9). When 0 I i.e. a attains the maximum, we have 6 ∈ | | d f c (c ) = f c (0) = a . ( ,a) 0 ( ,a) | | In this case we have

1 G(c, a)= g (c, a)= log f c (0) = log a = max c , a . 0 dn ( ,a) | | {| i| | |}

14 REIMI IROKAWA

Now we prove Proposition 4.9. First, it is easy to extend the result of Lemma 4.11 to the Ad−1,an analytification of K by continuity of each function. That is, the functions g(c,a)(z), gj(c, a), G(c, a) and log+ max c , a can be extended to a continuous function on Ad−1,an and we have {| i| | |} K (10) G(c, a) = log+ max c , a . {| j| | |} Take any convergent sequence λn in and let λ0 be the limit. Pd−{1 } M For the projective space K = Proj K[T0,...,Td−1], we have a natural open covering by affine d−1 space Ui i=0 , where Ui = Spec K[T0/Ti,...,Td−1/Ti]. Taking their analytification, we have an { } an Pd−1,an open covering Ui of K . { } an Now, take one of Ui such that an infinitely many λn belong to Ui and satisfy G(λn)= gi(λn), and replace λn by a subsequence with these properties. Without loss of generality, we may assume that U := U{ =} [c : : c : a : k] c = 0 , where k = 0 on P and k = 1 elsewhere and we 1 { 1 · · · d−2 | 1 6 } ∞ can take an affine coordinate of U as (c2/c1,...,cd−2/c1, a/c1, k/c1). Note that by assumption we have g (λ) = 0 for any n and g (λ ) as n . By (10), we have j 1 n →∞ →∞ log c (λ ) = F (λ )= G(λ ) log f (c (λ )) , | 1 n | n n ≥ | λn j n | whence (11) log c (λ ) log f (c (λ )) . | 1 n |≥ | λn j n | Note that 1 fλ(cj(λ)/c1(λ)) = d fλ(cj(λ)) c1(λ) since fλ(λ) is a homogeneous polynomial. It is enough to show fλ(cj(λn)/c1(λn)) 0 as n . From (11) we have → →∞ (1 d) log c (λ ) log f (c (λ )/c (λ )) . − | 1 n |≥ | λn j n 1 n | Taking n , we see that the left hand side tends to , which means f (c (λ )/c (λ )) 0. →∞ −∞ λ j n 1 n →

5. Normality vs. Passivity and J-stability vs. Activity Let V be a smooth and strictly K-analytic curve, f : V P1,an P1,an be an analytic family of rational functions of degree d, and c : V P1,an be a marked× critical→ point. As explained in Section 1, there are two ways to forumlate the→ activity in non-archimedean dynamics. The other approach by means of normality, is our main interest in this section. n As in Section 2.5, the family of functions cn(t) = ft (c(t)) n is said to be normal around t0 if for any subsequence c , there is an open neighborhood{ U of} t and a continuous function u on { nj } 0 U such that cnj pointwise converges to u. In this section, we compare the normality locus in this sense with the complement of the activity locus discussed in Section 3. In general, we have the following property:

Proposition 5.1. In the above notations, if c is attracted by an attracting periodic point w at t0, (f, c) is passive around t0. Proof. Replacing f by f n where n is the exact period of w and taking an Möbius transformation, we may assume that w(t) = 0 is a fixed point. Take an open neighborhood U of t0 small enough that 0 is attractive fixed point and c is converging to 0. Then clearly c (t) is normal as it { n } converges to a constant function 0. Consider a homogeneous lift Ft = (Pt,Qt) of ft and a suitable 2,an lift Ct : V A 0 of c(t). By definition, the activity measure on U is given by the Laplacian of the following→ function:\ { }

1 (n) (n) h(f,c)(t) = lim log max P (Ct) , Q (Ct) , n→∞ dn t t

ACTIVITY MEASURES OF DYNAMICAL SYSTEMS OVER NON-ARCHIMEDEAN FIELDS 15

(n) (n) where (Pt ,Qt ) is the n-th iteration of Ft. As c(t) is converging to 0, (n) n Qt (Ct) lim ft (c(t)) = lim = 0. n→∞ | | n→∞ P (n)(C ) t t (n) ( n) Therefore, for any n large enough, Pt (Ct) > Qt (Ct) . As we take Ft and Ct so that both (n) (n) | | | | Pt (Ct) and Qt (Ct) are not zero at the same time and neither of them has a pole, (n) (n) (n) log max Pt (Ct) , Qt (Ct) = log max Pt (Ct) is harmonic on U for every n large enough. Therefore, h(f,c)(t) is also harmonic on U and (f, c) is passive around t0. For families of polynomials considered in Section 4, we have the complete comparison as introduced in section 1: Proposition 5.2. Consider the same family as in the ﬁrst paragraph of Section 4.1 and assume that its boundedness locus (f,c) is non-empty and bounded. then the passivity locus coincides with the normality locus. M Proof. First, suppose t Supp µ and assume c (t) := f n(c(t)) is normal around t . We 0 ∈ (f,c) { n t }n 0 know that ∂ = Supp µ . As c is normal, there exists a subsequence c converging M(f,c) (f,c) { n} { nj }j to a continuous function φ around t0. Any open neighborhood U of t0 includes some point s such that c (s) as n . Therefore, by the continuity of φ we have φ(t ) = , which is n → ∞ → ∞ 0 ∞ contradiction as t0 (f,c) i.e. the orbit of t0 by cn must be bounded. For the contrary,∈ take M any point t Supp µ . Since Supp µ = ∂ , the point t 0 6∈ (f,c) (f,c) M(f,c) 0 belongs to either A1,an or ◦ , where ◦ is the (topological) interior of . In \ M(f,c) M(f,c) M(f,c) M(f,c) A1,an case t0 (f,c), there exists an open neighborhood U of t0 so that every point t of U has ∈ \ M 1,an unbounded orbit. Then, a family of functions cn(t) n on U is normal since cn(t) P on U. In case t ◦ , there exists a basic{ open} neighborhood U of t contained→∞∈ in . 0 ∈ M(f,c) 0 M(f,c) Since (f,c) is bounded, there is a real number M such that cn(t) < M for every t (f,c). In particular,M we have | | ∈ M ∞ c (U) P1,an t P1,an : t > M , n ⊂ \ ∈ | | n=1 [ where t P1,an : t > M is a generalized Berkovich open ball centered at the point . Then ∈ | | ∞ Proposition 2.17 asserts the normality of cn on U. { }

Corollary 5.3. Assume the residue characteristic of K is greater than d. In the above setting, every parameter t0 where f has an unstably indiﬀerent periodic point is in the activity locus of some critical point. Proof. We use the following proposition: Proposition 5.4 ([4], Theorem 1.2). If the residue characteristic of K is greater than d, then, for any rational function φ K(T ) and its attracting periodic point z, there exists some critical point c such that φn(c) is strictly∈ attracted by z.

As ft0 has an unstably indifferent periodic point, for any sufficiently small neighborhood U of t , there is a marked periodic point w(t) : U P1,an of f . Moreover, there is some tangent vector 0 → t ζ~ T (U) such that w(t) is attracting for every t represented by ζ~. Taking U smaller if necessary, ∈ t0 there exists a critical point c such that for every type I point t represented by ζ~, c(t) is strictly attracted by w(t). Meanwhile, There exists a tangent vector ξ~ T (U) such that w(t) is repelling ∈ t0 for all t represented by ξ~ i.e. for such t, c(t) is never attracted by w(t). Since any neighborhood 16 REIMI IROKAWA of t contains both of a point representing ζ~ and one representing ξ~, c := f n(c(t)) cannot be 0 { n t }n normal around t0. Remark 5.5. Even if the residue characteristic does not exceed d, the above corollary is still valid as long as ft is tame for every t. Actually, the exact condition that a critical point is attracted by an fixed point if 0 <λ< deg (f ) d | ζ,~v t | for all ζ t P1,an and every tangent vector ~v at ζ, where λ is a multiplier of the fixed point. ∈ { }× Here, we say a rational function φ is tame if for every z and its tangent vector ~v, degζ,~v(φ) cannot d be divided by the residue characteristic of K. When the family is tame, we have degζ,~v(ft) = 1, which means as long as a periodic point is attracting, there exists some critical point| attracted| by it. However, it is not true when a family has wild ramification; see the next section for a counter example.

6. Example: quadratic polynomials Now let us consider a concrete example. Assume the residue characteristic of K is 0 or greater 2 than 2. Let ft(z)= z + t, c(t) = 0, and consider the Mandelbrot set . For t 1, ft has a good M 2n−1 | |≤ reduction and the orbit of 0 is bounded. For t > 1, we have cn(t) = t . Indeed, ft(0) = t and for n> 0 we have | | | | | | | | | | n−2 n−1 c (t) = max(( t 2 )2, t )= t 2 | n | | | | | | | since t 2 > t > 1. This implies the orbit is unbounded. Hence = (0, 1) and the boundary | | | | M D ∂ = ζ0,1. Since the boundary consists of a single point, the activity measure should be a Dirac M n mass at the Gauss point ζ0,1. This is actually true because cn(t)= ft (c(t)) has good reduction for every n as a polynomial. We have the following fact: Proposition 6.1 ([18], Th´eor`eme E). A rational map φ of degree at least 2 has good reduction if and only if the canonial measure of φ is a Dirac mass δζ0,1 at the Gauss point.

Hence, the canonical measure must be a Dirac mass at the Gauss point δζ0,1 , so ∗ n cnδζ0,1 = 2 δζ0,1 , from which we have µ =µ ˆ = δζ0,1 . 1 On the other hand, the Julia set of ft with t > 1 is a Cantor set in the annulus z P (k) z = t while that of f with t 1 is a singleton| | ζ since f has good reduction.{ ∈ As a| | | | |} t | | ≤ { 0,1} t consequence a drastic bifurcation occurs at the point ζ0,1. Also [29] shows that the bifurcation locus equals the set of parameters where there exist some unstably indefferent periodic point or some repelling periodic point with multiplicity greater than 1. The point ζ0,1 corresponds to the one with unstably indefferent periodic point. Indeed, there is only one attracting periodic point z0 of ft: z0 = 1 √1 4t. Then, this is unstably indefferent if t = ζ0,1 since the absolute value of this is the following:− − t if t 1, and 1 √1 4t = | | | |≤ − − t 1/2 otherwise. ( | |

Remark 6.2. In case a family f has wild ramification for some t, the activity of critical points cannot describe the J-stability. We can easily see this in the same example as above but the residue characteristic is 2. The proof is straightforward: 2 Proposition 6.3. Assume that the characteristic of the residue field of K is 2. Set ft(z)= z + t, c(t) = 0 and c (t)= f n(c(t)) = f n(0). Denote by the Mandelbrot set as above and (f) be the n t t M B set of points in the parameter space where ft has unstably indifferent cycle. Then, ∂ = ζ , while M 0,1 ACTIVITY MEASURES OF DYNAMICAL SYSTEMS OVER NON-ARCHIMEDEAN FIELDS 17

(f)= ζ . B 0,4 In this case, the above two sets do not coincide. The argument which shows the Mandelbrot set is (0, 1) is valid in this case, too. Only thing D to consider here is about (f). As we change the coordinate as z z √t, we have the family B 7→ − g (z) = z 2√tz . This has good reduction if t 4. Hence, when t 4, the polynomial { t − } | | ≤ | | ≤ ft has potentially good reduction. In this case, the Julia set J(ft) moves continuously, so this should be J-stable even though this motion is out of the scope of the definition of J-stability in [29] since the stability is defined only if the Julia set has some classical point. For this case, the same drastic bifurcation described above occurs at ζ0,4, instead of ζ0,1, different from the boundary of the Mandelbrot set. The problem is from the constant ε in Theorem 5.4. In this case, we have ε = max(1, 2 2) = 1/4 < 1. | | To associate attracting periodic point to the behavior of critical points, we need to assume σ = 1. Remark 6.4. In [28], it has been discussed that the non-existence of unstably indifferent cycle is not enough for the J-stablity to hold; in fact, one has to add the non-existence of type 1 repelling point with multiplicity greater than 1. However, our activity measure cannot detect over which parameter there exists a type 1 repelling point with multiplicity greater greater than 1, since this has nothing to do with the asymptotic behavior of the critical point. The bifurcation of this kind occurs, for instance, at every point of the open segment (ζ0,1, ) in the above example, on which the critical point 0 escapes to . In this case, the Julia set is∞ always homeomorphic to a Cantor set, but there is no analytic motion∞ of repelling-periodic points (for details, see [28]). Our study via the activity measure, therefore, cannot detect the bifurcation phenomena of this type. However, we think it is still useful in studying bifurcations that changes the shape of Julia sets. References [1] Baker, Matthew, and Robert S. Rumely. “Potential theory and dynamics on the Berkovich projective line.” No. 159. American Mathematical Soc., 2010. [2] Baker, Matthew, and Laura DeMarco. “Preperiodic points and unlikely intersections.” Duke Mathematical Journal 159.1 (2011): 1-29. [3] Baker, Matthew, and Laura De Marco. “Special curves and postcritically finite polynomials.” Forum of Mathe- matics, Pi. Vol. 1. Cambridge University Press, 2013. [4] Benedetto, Robert, et al. “Attracting cycles in p-adic dynamics and height bounds for postcritically finite maps.” Duke Mathematical Journal 163.13 (2014): 2325-2356. [5] Berkovich, Vladimir G. “Spectral theory and analytic geometry over non-Archimedean fields.” No. 33. American Mathematical Soc., 2012. [6] Berteloot, Fran¸cois. “Bifurcation currents in holomorphic families of rational maps.” Pluripotential theory. Springer, Berlin, Heidelberg, 2013. 1-93. [7] Brolin, Hans. “Invariant sets under iteration of rational functions.” Ark. Mat., 6:103-144, 1965. [8] Chambert-Loir, Antoine, and Antoine Ducros. ”Formes différentielles réelles et courants sur les espaces de Berkovich.” arXiv preprint arXiv:1204.6277 (2012). [9] DeMarco, L. (2001). Dynamics of rational maps: A current of the bifucation locus. Mathematical Research Letters, 8(1-2), 57-66. https://doi.org/10.4310/MRL.2001.v8.n1.a7 [10] DeMarco, Laura. “Dynamics of rational maps: Lyapunov exponents, bifurcations, and capacity.” Mathematische Annalen 326.1 (2003): 43-73. [11] DeMarco, Laura. “Bifurcations, intersections, and heights.” Algebra & Number Theory 10.5 (2016): 1031-1056. [12] Dujardin, Romain, and Charles Favre. “Distribution of rational maps with a preperiodic critical point.” American journal of mathematics 130.4 (2008): 979-1032. [13] Dinh, Tien-Cuong, and Nessim Sibony. “Distribution des valeurs de transformations méromorphes et applications.” Commentarii Mathematici Helvetici 81.1 (2006): 221-258. [14] Dujardin, Romain. “Bifurcation currents and equidistribution on parameter space.” arXiv preprint 1111.3989 (2011). [15] Favre, Charles, Jan Kiwi, and Eugenio Trucco. “A non-Archimedean Montel’s theorem.” Compositio Mathemat- ica 148.3 (2012): 966-990. [16] Favre, Charles, and Thomas Gauthier. “Distribution of postcritically finite polynomials.” Israel Journal of Math- ematics 209.1 (2015): 235-292. 18 REIMI IROKAWA

[17] Favre, Charles, and Thomas Gauthier. “Classification of special curves in the space of cubic polynomials.” International Mathematics Research Notices 2018.2 (2016): 362-411. [18] Favre, Charles, and Juan Rivera-Letelier. “Théorème d’équidistribution de Brolin en dynamique p-adique.” Comptes Rendus Mathematique 339.4 (2004): 271-276. [19] Favre, Charles, and Juan Rivera-Letelier. ”Théorie ergodique des fractions rationnelles sur un corps ul- tramétrique.” Proceedings of the London Mathematical Society 100.1, 2010: 116-154. [20] Ghioca, Dragos, Hexi Ye. “The dynamical André-Oort conjecture for cubic polynomials.” Intern. Math. Res. Not.(IMRN) (2018). [21] Jonsson, Mattias. “Dynamics on Berkovich spaces in low dimensions.” Berkovich spaces and applications. Springer, Cham, 2015. 205-366. [22] LEE, JUNGHUN. ”J-stability of expanding maps in non-Archimedean dynamics.” Ergodic Theory and Dynam- ical Systems (2017): 1-18. [23] Lyubich, M. Yu. The measure of maximal entropy of a rational endomorphism of a Riemann sphere. Funktsional. Anal. i Prilozhen., 16:78-79, 1982. [24] Lyubich, M. Yu. “Some typical properties of the dynamics of rational maps.” Russian Mathematical Surveys 38.5 (1983): 154-155. [25] Mañé, Ricardo, Paulo Sad, and Dennis Sullivan. “On the dynamics of rational maps.” Annales scientifiques de l’Ecole´ Normale Supérieure. Vol. 16. No. 2. Elsevier, 1983. [26] McMullen, Curtis T. “Complex Dynamics and Renormalization” (AM-135). Vol. 135. Princeton University Press, 2016. [27] Rivera Letelier, Juan. “Dynamique des fractions rationnelles sur des corps locaux,” Diss. Paris 11, 2000. [28] Silverman, Joseph H. “The arithmetic of dynamical systems.” Vol. 241. Springer Science & Business Media, 2007. [29] Silverman, Thomas. “A non-archimedean λ-lemma.” arXiv preprint arXiv:1712.01372, 2017. [30] Thuillier, Amaury. “Théorie du potentiel sur les courbes en gèométrie analytique non archimédienne. applications àla théorie d’arakelov.” Diss. UniversitéRennes 1, 2005.

Graduate School of Science, Tokyo Insitute of Technology Email address: [email protected]