lt;/a> exists, thus the <a href="http://en.wikipedia.org/wiki/axiom_of_constructibility" class="extiw" title="wikipedia:axiom of constructibility">axiom of constructibility</a> ($V=L$) fails. - In fact, $x^{\sharp}$ exists for every $x\in\mathbb{R}$ (thus $V\neq L\[x\]$). - The strong [](Partition%20property.md), $\omega_1\rightarrow(\omega_1)^{\omega_1}_2$, holds. In fact, $\omega_1\rightarrow(\omega_1)^{\omega_1}_{2^{\aleph_0}}$ and $\omega_1\rightarrow(\omega_1)^{\omega_1}_\alpha$ for every $\alpha<\omega_1$. - If there is a surjection $\mathbb{R}\to\alpha$, then there is surjection $\mathbb{R}\to\mathcal{P}(\alpha)$. - Equivalently, if $\alpha\preceq 2^{\aleph_0}$ then $2^\alpha\preceq 2^{\aleph_0}$ - <a href="http://en.wikipedia.org/wiki/Hall%27s_marriage_theorem" class="extiw" title="wikipedia:Hall's marriage theorem">Hall's marriage theorem</a> fails for infinite graphs. For example there is there is a 2-regular bipartite graph on $\mathbb{R}$ with no perfect matching. - There is no <a href="http://en.wikipedia.org/wiki/Basis_(linear_algebra)#Related_notions" class="extiw" title="wikipedia:Basis (linear algebra)">Hamel basis</a> of $\mathbb{R}$ over $\mathbb{Q}$. Let $\Theta$ be the supremum of the ordinals that $\mathbb{R}$ can be mapped onto: $\Theta=\mathrm{sup}\{\alpha\in\mathrm{Ord}:\alpha\preceq 2^{\aleph_0}\}$. Under $\text{AC}$ this is just $(2^{\aleph_0})^{+}$ but under $\text{AD}$ it is a limit cardinal, in fact an aleph fixed point, and $\text{DC}$ implies it has uncountable cofinality. In $L(\mathbb{R})$ it is also regular and thus <a href="Weakly_inaccessible" class="mw-redirect" title="Weakly inaccessible">weakly inaccessible</a>. It is conjectured that under $\text{AD}$ the cofinality function is nondecreasing on singular cardinals below $\Theta$. It is worth noting that, under $\text{AD}$, $\Theta$ is the least ordinal *incomparable* with $\mathbb{R}$, i.e. such that $\Theta\not\leq 2^{\aleph_0}$ yet $2^{\aleph_0}\not\leq\Theta$ and $\Theta\not\preceq 2^{\aleph_0}$ yet $2^{\aleph_0}\not\preceq\Theta$ Under $\text{ZF+AD}$, the cardinal $\omega_1$ is <$\Theta$-[](Strongly%20compact.md) in that for all $\lambda<\Theta$, every $\sigma$-complete filter on $\lambda$ extends to a ultrafilter on $\lambda$ (and every ultrafilter is itself $\sigma$-complete under $\text{AD}$.) ## Determinacy of $L(\mathbb{R})$ *See also: [](Constructible%20universe.md)* Recall that a formula $\varphi$ is $\Delta_0$ if and only if it only contains bounded quantifiers (i.e. $(\forall x\in y)$ and $(\exists x\in y)$). Let $\text{def}(X)=\{Y\subset X : Y$ is first-order definable by a $\Delta_0$ formula with parameters only from $X\cup\{X\}\}$. Then let: - $L_0(X)=\text{TC}(\{X\})$ - $L_{\alpha+1}(X)=\text{def}(L_\alpha(X))$ - $L_\lambda(X)=\bigcup_{\alpha<\lambda}L_\alpha(X)$ for limit $\lambda$ - $L(X)=\bigcup_{\alpha\in \text{Ord}}L_\alpha(X)$ where $\text{TC}({X})$ is the smallest transitive set containing $X$, the elements of $X$, the elements of the elements of $X$, and so on. $L(X)$is always a model of $\text{ZF}$, but not necessarily of the axiom of choice. $L(X,Y)$ is used as a shortcut for $L(\{X,Y\})$. $L(X,\mathbb{R})$ with $X\subset\mathbb{R}$ is different from $L(\mathbb{R})$ whenever $X$ is not constructible from the reals, i.e. $X\not\in L(\mathbb{R})$ (if any such set exists; it is consistent with $\text{ZF+AD}$ that they do not). $L(\mathbb{R})$-determinacy, also known as $\text{AD}^{L(\mathbb{R})}$ {% cite Koellner2010 %} or *quasi-projective determinacy* {% cite Maddy1988 %} is the assertion that every set of reals in $L(\mathbb{R})$ is determined. Equivalently, "$L(\mathbb{R})$ is a model of $\text{ZF+AD}quot;. $\text{AD}^{L(\mathbb{R})}$ appears to be a very "natural" statement in that, empirically, every natural extension of $\text{ZFC}$ (i.e. not made specifically to contradict this) that is not proved consistent by $\text{AD}$ seems to imply $\text{AD}^{L(\mathbb{R})}$ or some weaker form of determinacy. {% cite Larson2013 %} This is often considered to be an argument toward the "truth" of $\text{AD}^{L(\mathbb{R})}$. Assuming $\text{ZF+DC+}V=L(\mathbb{R})$, $\text{AD}$ follows from three of its consequences: {% cite Larson2013 %} 1. Every set of reals is Lebesgue measurable. 2. Every set of reals has the Baire property. 3. Every $\Sigma^1_2$ set of reals can be uniformized. In $L(\mathbb{R})$, the axiom of determinacy is equivalent to the axiom of Turing determinacy {% cite Larson2013 %}, i.e. the assertion that payoff sets closed under <a href="http://en.wikipedia.org/wiki/Turing_equivalence" class="extiw" title="wikipedia:Turing equivalence">Turing equivalence</a> are determined. Busche and Schindler showed that, if there is a model of $\text{ZF}$ in wich every uncountable cardinal is singular (thus has cofinality $\aleph_0$), then the axiom of determinacy holds in the $L(\mathbb{R})$ of some [](Forcing.md) of $\text{HOD}$ {% cite Larson2013 %}. This notably follows from the existence of a proper class of [](Strongly%20compact.md) cardinals. Assume that there is $\omega_1$-[](Filter.md) over $\omega_1$; then $\text{AD}^{L(\mathbb{R})}$ holds. {% cite Kanamori2009 %} This result is due to Woodin. The following holds in $L(\mathbb{R})$ assuming $\text{AD}^{L(\mathbb{R})}$: {% cite Koellner2010 Jackson2015 %} - Every uncountable cardinal <$\Theta$ is [Jónsson](Jonsson.md "Jonsson"), also if it is regular or has cofinality $\omega$ then it is [Rowbottom](Rowbottom.md "Rowbottom"). - Every regular cardinal <$\Theta$ is [measurable](Measurable.md "Measurable") (note that $2^{\aleph_0}\not\leq\Theta$), also $\Theta$ is a limit of measurable cardinals. - $\Theta$ is weakly hyper-[Mahlo](Mahlo.md "Mahlo") (and thus weakly hyper-inaccessible), but it is not [[weakly compact]]. - $\Theta$ is Woodin in the model $\text{HOD}^{L(\mathbb{R})}$. There exists many strengthnings and variants of this result. - $\omega_1$ is <$\Theta$-supercompact in that for every $\lambda<\Theta$ there is a normal ultrafilter on $P_{\omega_1}(\lambda)$. This is a strengthening of the above result that $\omega_1$ is <$\Theta$-strongly compact under $\text{AD}$ alone. ## Axiom of projective determinacy *Main article: [](Projective.md#Projective_determinacy)* ## Axiom of real determinacy The **axiom of real determinacy** ($\text{AD}_\mathbb{R}$) is the assertion that if payoff sets contains real numbers instead of natural numbers, then every payoff set is still determined. This is strictly stronger than $\text{AD}$, and $\text{ZF+AD}_\mathbb{R}$ proves $\text{ZF+AD}$ consistent. $\text{AD}_\mathbb{R}$ is equivalent (over $\text{ZF}$) to $\text{AD}$ plus the <a href="http://en.wikipedia.org/wiki/Uniformization_(set_theory)" class="extiw" title="wikipedia:Uniformization (set theory)">axiom of uniformization</a> (which is false in $L(\mathbb{R})$). $\text{AD}_\mathbb{R}$ is also equivalent to determinacy for games of length $\omega^2$. In fact, $\text{AD}_\mathbb{R}$ is equivalent to the assertion that every game of bounded countable length is determined. It is however possible to show (in $\text{ZF}$) that there are non-determined games of length $\aleph_1$. Solovay showed that $\text{ZF+AD}_\mathbb{R}+quot;$\Theta$ has uncountable cofinality" (which follows from $\text{ZF+AD}_\mathbb{R}\text{+DC}$) proves $\text{ZF+AD}_\mathbb{R}$ consistent; it is therefore consistent with $\text{ZF+AD}_\mathbb{R}$ that $\Theta$ has cofinality $\omega$ and that $\text{DC}$ is false. {% cite Larson2013 %} Steel showed that under $\text{AD}_\mathbb{R}$, in a [](Forcing.md) there is a proper class model of $\text{ZFC}$ in which there exists a cardinal $\delta$ of cofinality $\aleph_0$ which is a limit of Woodin cardinals and <$\delta$-strong cardinals. {% cite Larson2013 %} Under $\text{AD}_\mathbb{R}$, $\omega_1$ is <$\Theta$-supercompact, i.e. for every ordinal $\gamma<\Theta$ there is a normal fine ultrafilter on the set of all subsets of $\gamma$ of size $\aleph_1$. $\text{AD}$ suffices for this result to hold in $L(\mathbb{R})$, but is not known to suffice for it to hold in $\text{V}$. {% cite Larson2013 %} A set $\Gamma\subset\mathcal{P}(\mathbb{R})$ is a *Wadge initial segment* of $\mathcal{P}(\mathbb{R})$ if for every $X\in\Gamma$, if $Y\leq_W X$ (i.e. $Y$ is <a href="http://en.wikipedia.org/wiki/Wadge_hierarchy" class="extiw" title="wikipedia:Wadge hierarchy">Wadge reducible</a> to $X$) then $Y\in\Gamma$. Under suitable large cardinal assumptions, there exists a Wadge initial segment $\Gamma\subset\mathcal{P}(\mathbb{R})$ such that $L(\Gamma,\mathbb{R})\models\text{AD}^{+}+\text{AD}_\mathbb{R}+\Gamma=\mathcal{P}(\mathbb{R})$ (see <a href="AD%2B" class="mw-redirect" title="AD+">AD+</a>). Furthermore, whenever $\mathcal{M}$ is an inner model such that $\mathbb{R}\subset\mathcal{M}$ and $\mathcal{M}\models\text{AD}^{+}+\text{AD}_\mathbb{R}$, one has $\Gamma\subset\mathcal{M}$. *(see the 'Read more' section)* ## $\text{AD}^+$ and models of determinacy *Main article: <a href="AD%2B" class="mw-redirect" title="AD+">$\text{AD}^+lt;/a>* ## Consistency strength of determinacy hypotheses The following theories are equiconsistent: {% cite Kanamori2009 Trang2016 %} - $\text{ZF+AD}$ - $\text{ZF+AD+DC}$ - $\text{ZFC+AD}^{L(\mathbb{R})}$ - $\text{ZFC+AD}^{\text{OD}(\mathbb{R})}$ - $\text{ZFC+}quot;the non-stationary ideal over $\omega_1$ is $\omega_1$-dense" - $\text{ZFC+}quot;there exists infinitely many [Woodin](Woodin.md "Woodin") cardinals" - $\text{ZF+DC+}quot;$\omega_1$ is $\mathcal{P}(\omega_1)$-[](Strongly%20compact.md)" - $\text{ZF+DC+}quot;$\omega_1$ is $\mathbb{R}$-strongly compact and $\Theta>\omega_2quot; - $\text{ZF+DC+}quot;$\omega_1$ is $\mathbb{R}$-strongly compact and $\omega_2$-strongly compact" - $\text{ZF+DC+}quot;$\omega_1$ is $\mathbb{R}$-strongly compact and Jensens's square principle fails for $\omega_1quot; Where $\text{DC}$ is the <a href="http://en.wikipedia.org/wiki/axiom_of_dependent_choice" class="extiw" title="wikipedia:axiom of dependent choice">axiom of dependent choice</a> and $\omega_1$ being $X$-strongly compact means that there exists a [](Filter.md) on the set of all subsets of $X$ of cardinality $\aleph_1$. Woodin proved that the theory $\text{ZF+DC+AD+}quot;$\omega_1$ is supercompact" is consistent relative to $\text{ZFC+}$ "there is a proper class of Woodin cardinals, each of which a limit of Woodin cardinals". <a href="Projective_determinacy" class="mw-redirect" title="Projective determinacy">Projective determinacy</a> is a little weaker: it is equiconsistent with $\text{ZFC}$ plus, for all n, an axiom saying "there are n Woodin cardinals". Since $\text{ZFC}$ can only use finitely many of its axioms, this axiom schema does not allow $\text{ZFC}$ to prove that there exists infinitely many Woodins, despite making it able to prove every particular instance of "there exists at least n Woodin cardinals". Koellner annd Woodin showed that the following theories are also equiconsistent: {% cite Koellner2010 %} - $\text{ZFC+}\Delta^1_2$-determinacy - $\text{ZFC+OD}$-determinacy - $\text{ZFC+}quot;there exists a Woodin cardinal" - $\text{ZFC+}quot;the nonstationary ideal on $\omega_1$ is $\omega_2$-saturated" And so are $\text{Z}_3+$lightface $\Delta^1_2$-determinacy and $\text{MK+}quot;$\text{Ord}$ is Woodin" where $\text{Z}_3$ is *third-order arithmetic* and $\text{MK}$ is <a href="Morse-Kelley_set_theory" class="mw-redirect" title="Morse-Kelley set theory">Morse-Kelley set theory</a>. It is also conjectured that $\text{Z}_2+\Delta^1_2$-determinacy and $\text{ZFC+}quot;$\text{Ord}$ is Woodin" are equiconsistent, where $\text{Z}_2$ is <a href="http://en.wikipedia.org/wiki/second-order_arithmetic" class="extiw" title="wikipedia:second-order arithmetic">second-order arithmetic</a> and "$\text{Ord}$ is Woodin" is expressed as an axiom scheme. Finally, Trang and Wilson proved that the following theories are equiconsistent: {% cite Trang2016 %} - $\text{ZF+DC+AD}_\mathbb{R}$ - $\text{ZF+DC+}quot;$\omega_1$ is $\mathcal{P}(\mathbb{R})$-strongly compact" - $\text{ZF+DC+}quot;$\omega_1$ is $\mathbb{R}$-strongly compact and $\Theta$ is singular" - $\text{ZF+DC+}quot;$\omega_1$ is $\mathbb{R}$-strongly compact and $\Theta$-strongly compact" As are the following theories: - $\text{ZF+AD}_\mathbb{R}$ - $\text{ZF+DC}_{\mathcal{P}(\omega_1)}+quot;$\omega_1$ is $\mathbb{R}$-strongly compact and $\Theta$ is singular" - $\text{ZFC+}quot;there is a cardinal $\lambda$ that is a limit of Woodin cardinals and <$\lambda$-strong cardinals". ## Read more - *"Is there a natural inner model of $\text{AD}_\mathbb{R}$?"* <a href="http://mathoverflow.net/questions/269241/is-there-a-natural-inner-model-of-ad-mathbbr/269690" class="external autonumber">[3]</a> - *"Limitations of determinacy hypotheses in ZFC"* <a href="http://mathoverflow.net/questions/271507/limitations-of-determinacy-hypotheses-in-zfc" class="external autonumber">[4]</a> - *"Counterintuitive consequences of the Axiom of Determinacy?"* <a href="https://mathoverflow.net/questions/129036/counterintuitive-consequences-of-the-axiom-of-determinacy" class="external autonumber">[5]</a>