Continuation from thread:
https://groups.google.com/d/msg/talk.origins/3ttASLL3nfI/EOqh_3m3AQAJ
Specifically Post:
https://groups.google.com/d/msg/talk.origins/3ttASLL3nfI/3rpls5dSCAAJ
Reason for continuation: Replies not coming through.
----------------
On Monday, April 9, 2018 at 1:55:03 PM UTC+1, Andre G. Isaak wrote:
> In article <
41827e5f-d1d9-4dd1...@googlegroups.com>,
> someone wrote:
>
> > On Friday, April 6, 2018 at 4:55:03 PM UTC+1, Andre G. Isaak wrote:
> > > In article <
dd63eb8f-ace9-4c9a...@googlegroups.com>,
> > > someone wrote:
> > >
> > > > On Friday, April 6, 2018 at 6:10:03 AM UTC+1, Andre G. Isaak wrote:
> > > > > In article <
337b0cc2-c3c8-4220...@googlegroups.com>,
> > > > > someone wrote:
> > > > >
> > > > > > On Thursday, April 5, 2018 at 9:40:04 PM UTC+1, Andre G. Isaak wrote:
> > > > > > > In article <
8641b387-cca9-4566...@googlegroups.com>,
> > > > > > > someone wrote:
> > > > > > >
> > > > > > > > On Thursday, April 5, 2018 at 7:05:04 PM UTC+1, Andre G. Isaak
> > > > > > > > wrote:
> > > > > > > > > In article
> > > > > > > > > <
2d11c48a-66a0-44de...@googlegroups.com>,
> > > > > > > > > someone wrote:
> > > > > > > > >
> > > > > > > > > > On Thursday, April 5, 2018 at 6:00:03 AM UTC+1, Andre G.
> > > > > > > > > > Isaak
> > > > > > > > > > wrote:
> > > > > > > > > > > In article
> > > > > > > > > > > <
58b0a5f0-f8ec-4283...@googlegroups.com>,
> > > > > > > > > > > someone wrote:
> > > > > > >
> > > > > > But the conversation has moved on from fine tuning to objective
> > > > > > bayesian
> > > > > > inference, because we seemed to be talking at cross purposes when
> > > > > > discussing
> > > > > > fine tuning.
> > > > >
> > > > > You have 'moved on' to talking about Bayesian inference. I am still
> > > > > trying to explain why fine-tuning is a non-argument. Bayesian inference
> > > > > doesn't help here because the problem is not sufficiently well-defined.
> > > > >
> > > >
> > > > It does not have to be well-defined, we can just go on the basis of the
> > > > information we have. If we have no reason to have expected a certain
> > > > result,
> > > > then assume all results were equally as likely until we have reason to
> > > > assume
> > > > otherwise. Sure what the bayesian inference was could change the more
> > > > information we have but that is bayesian inference for you.
> > >
> > > So please demonstrate to me how you would apply Bayesian analysis to the
> > > following problem:
> > >
> > > I present you with a box which I assure you has certain TARDIS-like
> > > properties (i.e. you can't tell how big the inside of the box is by
> > > looking at the outside of the box -- I gather you are posting from the
> > > UK so I will assume this is a familiar reference).
> > >
> > > What are the odds of drawing three black cubes from the box in a row?
> > >
> >
> > I think you need to think of it more as comparing one hypothesis that gives
> > reason to expect what is drawn from the box will be black cubes (analogous to
> > theism)
>
> Ummm. Why would theism lead you to believe black cubes will be drawn?
> Note that this example is unrelated to the constants example. The black
> cubes mentioned above don't stand in for values of physical constants.
> They're simply black cubes.
>
I never stated that theism would lead anyone to believe black cubes to be drawn. I was making the black cube situation analogous to the physical constants situation and explaining how the analogy was working.
> > and another hypothesis which gives no reason to give any expectation
> > of the colour or the shape (analogous to physicalism) and then considering
> > which hypothesis given objective bayesian inference best explains drawing
> > three black cubes.
>
> The entire point this example makes is that you need *something* to base
> initial probabilities on before you can do bayesian analysis. In the
> above scenario, I've given no information whatsoever. I haven't said
> whether the box contains anything at all, let alone cubes. There's
> simply no way to do any meaningful probability calculations in absence
> of at least some information.
>
> Where the physical constants are concerned, the only thing we currently
> know is what the values are in this universe. But to calculate the
> probability of those values occurring would require some knowledge of
> the range of possible values, and we currently don't have that at all,
> so any numbers are simply being pulled out of people's asses. If you
> want to claim the current values are highly unlikely, just pick a big
> range. If you want to claim they're very likely, pick a very narrow
> range. Neither of those assumptions is anymore warranted than the other,
> which makes the probabilities garbage. That's true whether your a
> theist, a physicalist, an existentialist, a solipsist, a cubist, or any
> other -ist you choose.
>
In the black cube analogy *I gave* where you know a result (drawing 3 black cubes), that is analogous to the physics constants situation. And in that analogy there is enough information to give use objective bayesian inference to favour the analysis the hypothesis that expected what would be drawn from the box would be black cubes, even if no more drawings from the box could be done.
> > The issue is slightly more complicated however with the physics constants
> > situation, because physicalism allows for a multiverse type argument which
> > can then be combined with an anthropomorphic argument in which it is
> > analogous to saying there were lots of results which weren't just black
> > cubes, but we should only expect to observe the results that are.
>
> Multiverses don't help either. We're still left with the same
> fundamental problem: we have absolutely zero knowledge of the range of
> possible values. That's true whether there's one universe, a countable
> infinity of universes, or a nondenumerable number of universes.
>
The multiverse helps in the same way as the hypothesis that there are multiple drawings from the box, but we only observe the ones where black cubes are pulled, helps in the black cube analogy I gave. You presumably understood why in the black cube analogy why that hypothesis and the hypothesis that only black cubes would be drawn, would be favoured by objective bayesian inference over a hypotheses that gave no reason to have expected any particular colour or shape. It has to do with the a priori expectation for such a result given the hypothesis.
The hypothesis that reality is a physical one and gives no reason to expect the physical constants to be in the range that allows complex chemistry, whereas the hypothesis that reality is a physical one and that there is a large multiverse in which the physical constant values do would. Surely it is not hard for you to understand that the second hypothesis offers and explanation for why they are in the range that they are observed as being in but the first does not.
>
> > > > > > And objective bayesian inference is about the extent the results
> > > > > > were expected given the hypothesis. If the hypothesis implied the
> > > > > > results
> > > > > > then that would be different, but without a hypothesis that does, we
> > > > > > have
> > > > > > to
> > > > > > evaluate the hypotheses that we do have. So you could have a
> > > > > > hypotheses
> > > > > > that
> > > > > > reality is a physical one in which the constants are constant
> > > > > > throughout
> > > > > > the
> > > > > > universe and cannot because of physical reasons vary. But that does
> > > > > > not
> > > > > > lead
> > > > > > you to expect the results that were discovered.
> > > > >
> > > > > Nor does physics provide an explanation for why the distance between
> > > > > Baltimore and DC is what it is. So what?
> > > > >
> > > >
> > > > In this case it is not really to do with physics, but metaphysics. It is
> > > > doubtful that the hypothesis of a physicalist universe/multiverse with
> > > > varying physics constant values could ever be tested vs the hypothesis of
> > > > a
> > > > physicalist single universe in which the physics constants are constant
> > > > throughout could ever be tested.
> > >
> > > This is precisely why I give little credence to the multiverse
> > > hypothesis. A hypothesis which cannot in principle be tested is not a
> > > hypothesis at all.
> > >
> >
> > That is simply wrong. There are metaphysical hypotheses.
>
> Untesstable 'metaphysical' hypotheses amount to no more than
> navel-gazing. Unless a hypothesis makes some meaningful claim about the
> universe, it's of no interest to me.
>
So the claim of physicalism are of no interest to you because it is an untestable metaphysical hypothesis? I thought you were a physicalist.
> > And even if you
> > advocate using Occam's razor to favour the simplest to decide which to
> > favour, you are often ending up favouring a metaphysical hypotheses.
>
> That's a misinterpretation of occam's razor. Occam's razor is simply a
> heuristic. It isn't a principle governing how the universe works.
>
How is it a misinterpretation? It does work by favouring a simpler metaphysical hypothesis over a more complex one.
> > Sure you could go for the approach in which you pretend only to understand
>
> I'd forgotten how obnoxious you can be. Please do not accuse your
> interlocutors of 'pretending' not to understand something. I find many
> of your views entirely ludicrous, but I am nonetheless willing to assume
> that they are genuinely your views. Please do me the same courtesy.
>
>
> > what is meant if it means something in terms of experimental results. If that
> > is the type of direction you favour could you mention it now, because it
> > would mean that you cannot understand what is meant by a multiverse, because
> > you do not understand what experimental result is being referred to. Though I
> > guess you giving little credence to the hypothesis of a multiverse (you did
> > realise it was just a hypothesis) means you did understand it, and are not
> > planning on retreating to such a position (though (as I remember it) you have
> > before, requiring operational definitions for experience for example). Though
> > if you were planning on possibly doing so, could you make it clear at this
> > point whether that is your position?
>
> I've always been very clear on my position in this regards. Claims that
> make absolutely no testable claims aren't worth serious consideration. I
> have no issue with people pondering them or using them in works of
> fiction, but they have no place in legitimate scientific arguments (or
> philosophical arguments, for that matter).
>
So would you understand what was being considered if people were to consider whether a robot was experiencing or not, but that you would not find it interesting unless there was a test to tell?
> > > > But objective bayesian inference gives us
> > > > reason to favour the universe with varying constant values hypothesis
> > > > over
> > > > the hypothesis that the values are constant.
> > >
> > > <snip>
> > >
> > > > > > As a side issue, even if cosmological constants were determined by
> > > > > > the
> > > > > > laws
> > > > > > of physics, physicalism would have no reason to have expected those
> > > > > > laws.
> > > > > >
> > > > >
> > > > > Physics seeks to discover the laws of nature objectively. Why would one
> > > > > expect it to start with any expectations?
> > > > >
> > > >
> > > > Physicalism is not the same as physics. From a theist perspective it is
> > > > easy
> > > > to see why the physics would be expected to allow forms which were
> > > > suitable
> > > > for influencing morally.
> > >
> > > What does it mean for a form to 'influence morality'? And as I stated in
> > > a previous post, I fail to see how this in any way leads to conclusions
> > > about what the universe would look like.
> > >
> >
> > From a theist perspective, I would expect that if I was to be given the
> > experience of having a form that I would be able to influence the behaviour
> > of the form based on the experience in such a way that it would seem
> > plausible to me that such behaviour would influence the experience of others.
>
> That's still no more clear than the original.
>
If there is some part of what I wrote there that you are not clear about, then point it out.
> >
> > > > I noticed you did not answer the question. Were you suggesting hypothesis
> > > > 1
> > > > or 2 for the bayesian inference, or were you rejecting bayesian inference
> > > > as
> > > > a best practise approach when deciding between two hypotheses?
> > >
> > > I do not believe that the problem at hand is one which Bayesian analysis
> > > can help us answer. If you can provide a good answer to the question
> > > posed at the beginning of this post, I will reconsider.
> > >
> > >
> >
> > That is twice now you avoided answering the question.
>
> I quite clearly did answer it. If it was not clear from my response, I
> was choosing your option 3 (i.e. I was rejecting Bayesian inference as a
> productive way of approaching this problem).
>
> Also, considering you snipped the entire second portion of my post at
> the beginning of this thread (the half containing my more salient
> points), I wouldn't complain about people not answering every one of
> your questions.
>
The reason for snipping was that there was no point in discussing things which relied on you understanding the how objective bayesian inference could be applied to the problem, when you still do not seem to be able to. And the third point, was not you being unable to see how objective bayesian inference could be applied to the problem, it was about rejecting objective bayesian inference as being best practise when deciding between two hypotheses.
Consider the following hypotheses:
1) Given the known laws of physics the mass values of the up quark down quark and electron will be within a range which allows lifeforms.
2) Given the known laws of physics there is no prediction for the values of the up quark down quark and electron other than they will be above 0 and below the Planck Mass.
With this situation given the actual mass values one could use objective bayesian inference to evaluate likelihood of those hypotheses providing a result.
The evaluation simply takes into account the results the hypothesis constrains the outcome to, and evaluates each result within the range as equally as likely (because it is objective bayesian inference we are discussing, not subjective bayesian inference) for the hypothesis and those outside the range have a probability of 0 given the hypothesis. So it is not a case of the probability evaluation being subjectively made up or "being pulled out of people's asses" as you seem fond of saying. Because regardless of how many decimal places the results are considered up to, the ratio of relative likelihood between the two hypotheses will remain pretty much the same and thus will be an objective result of applying objective bayesian inference. So it does not matter who does the evaluation or how many do it, they should all come to the same conclusion using the method.
>
> > > > > > > 'Pick a number between 1 and 1,000,000' is constrained.'
> > > > > > >
> > > > > > > 'Pick a number' is not. If your person in a cubicle picks a number
> > > > > > > and
> > > > > > > your magic cubicle picks a number, you can't calculate the odds of
> > > > > > > the
> > > > > > > numbers matching without specifying some sort of range as above.
> > > > > > >
> > > > > >
> > > > > > But what you can do is if there is no constraint then add one if it
> > > > > > seems
> > > > > > reasonable to do so. So many add Planck's mass as a constraint to
> > > > > > considered
> > > > > > mass values for cosmological constants.
> > > > >
> > > > > This is entirely non-sensical. For starters, the Planck Mass is
> > > > > *derived* from the values of the other physical constants, so if you
> > > > > allow those constants to vary, the planck mass varies with them.
> > > > >
> > > >
> > > > But it is not derived from other mass constants. So considering
> > > > variations of
> > > > those does not change it.
> > >
> > > But we were considering the possibility of physical constants varying,
> > > and that would in turn mean that the planck mass cannot be assumed to be
> > > constant across universes ex hypothesi.
> > >
> >
> > It does not matter whether the Planck Mass changes across universes. The
> > Planck Mass can change, and the ranges of mass considered will vary depending
> > on what the Planck Mass is in that subset of considerations (the universes
> > with that Planck Mass). All cosmologists are doing here is considering the
> > masses given the subset with the Planck Mass that we have. Though actually
> > the example I gave the range considered did not go even near to the Planck
> > Mass. It restricted the consideration of the masses of the electron, up quark
> > and down quark, to the mass of the top quark. And even with that quite large
> > (given the range) restriction of consideration, the odds of complex chemistry
> > given our current knowledge appear to be in the region of roughly 1 in 5
> > trillion. Obviously if you considered the full range (up to Planck Mass given
> > the constants that it is based on in this "universe") then it being that
> > likely would appear crazily optimistic.
> >
> > > > > And, when talking about the masses of quarks, the Planck Mass is simply
> > > > > an absurd value to choose as a constraint (the planck mass is something
> > > > > like 20 orders of magnitude more massive than a hydrogen atom -- hardly
> > > > > a reasonable starting point for thinking about subatomic particles).
> > > > >
> > > > > The bottom line is that any range constraints you choose here are
> > > > > basically being pulled out of your ass. Which makes any kind of
> > > > > probabilistic arguments (Bayesian inference, for example) entirely
> > > > > meaningless.
> > > > >
> > > >
> > > > The Planck mass is a natural limit to the range considered, as our
> > > > physics
> > > > stops there.
> > >
> > > ??? What do you mean our physics *stops* there ???
> > >
> > > I think you are confused about what a Planck Mass is.
> > >
> > > <snip>
> > >
> >
> > It is the mass where a particle is predicted to become its own black hole and
>
> No. It isn't. It is simply a unit of measurement.
>
> There is no mass at which something becomes a black hole. There is a
> mass at which something of a particular diameter (or wavelength when
> dealing with quantum phenomena) becomes a black hole.
>
> You're confusing the planck mass with the (hypothetical) planck
> particle, which has a mass equal to the planck mass and a diameter equal
> to the planck length. AFAIK, no one has seriously proposed the existence
> of such things. They're more of a pedagogical exercise than anything
> else.
>
Well we were considering electrons and quarks, which are fundamental particles, and as I understand it they tend to be considered as point particles (or in string theory a string with Planck length). So as I understand it, if their mass was to be equal to or greater then the Planck mass they are predicted to become black holes.
> > we do not currently have a theory of quantum gravity, so it currently sits at
> > a point which is at the limit of our physics. So perhaps the phrase "stops
> > there" was confusing as it did not indicate that I meant, and if a person was
> > to be pedantic, then they might have pointed out that our physics need not
> > stop there in the future. Though if you had realised the current limitation
> > regarding then why the multiple question marks. Would you mind explaining?
> > (Were you just being needlessly pedantic?)
>
> In this case the question marks indicated that I had absolutely no idea
> where you could have even gotten such a strange idea from.
>
> > > > Complex chemistry is the common focus, because it assumed that there
> > > > could be
> > > > no lifeforms without it. So while we are not sure what types of complex
> > > > chemistry would allow lifeforms, it is assumed that the constant
> > > > parameter
> > > > space which allows lifeforms will be a subset of the constant parameter
> > > > space
> > > > which allows complex chemistry.
> > >
> > > But then this simply leaves us with the question of why life is what you
> > > are focussing on rather than, say, black holes.
> > >
> >
> > Because theism gives no reason to expect black holes. But it does give reason
> > to expect that reality would not be one in which forms existed but did not
> > interact with each other in ways which could be considered moral. And the
> > point is to pick features which are useful in deciding which features allow a
> > distinction between which hypotheses should be favoured given objective
> > bayesian inference. Do you feel that I might have vaguely mentioned this
> > usefulness in distinguishing between hypotheses before?
>
> Obviously you've mentioned it. That doesn't make it in any way correct.
>
So I had answered the question already.
> Whatever strange theological theory you've come up with is a theory that
> you developed in this universe, so the fact that (you think) it is
> consistent with this universe is entirely unsurprising.
>
> What would be more surprising would be a theory which is entirely
> inconsistent with this universe (say, for example, if my view entailed
> there could be no elements lighter than the actinides. If I had such a
> view I might use it to argue that this universe is massively unlikely,
> but a far more reasonably conclusion would be that my view is simply
> wrong.
>
> Bottom line: any argument which rests on the probability of the universe
> being the way it is is rubbish for the simple reason that the
> probability calculations themselves are necessarily rubbish.
>
> >
> > > <snip>
> > >
> > > > I personally am not using the amazingly narrow range of physics constant
> > > > parameter space which allows for complex chemistry to argue for a fine
> > > > tuner.
> > > > After all what is wrong with the physicalist multiverse type replies?
> > >
> > > It is unnecessary. I don't feel a need to find a solution to a
> > > particular problem until I have been shown that the problem actually
> > > exists.
> > >
> >
> > In the cubicle Type 1 and cubicle Type 2 example, what indicates there is
> > there a problem with providing an answer that the cubicle is of Type 1? Do
> > you think there is no problem with the answer that the cubicle is of Type 2?
> > I hope you do not resort to stating that the situation is not the same while
> > avoiding answering the question.
> >
> > > > Do you accept that there is an apparent coincidence in that so many of
> > > > the
> > > > physics constants seem to be in the narrow value range that allows for
> > > > complex chemistry (it is not like 1 or 2 just happen to be in that
> > > > range)?
> > >
> > > No. I do not. At least no more than I accept the striking coincidence
> > > that a lakebed has exactly the same shape as the lake which it contains.
> > >
> >
> > But if there was a lake found on an uninhabited planet that was in the shape
> > of the sentence "God exists" and you were to claim that it was not by design,
> > then would you not accept a coincidence there? If that is not analogous to
> > what you were saying then perhaps explain why.
>
> No, that's not even remotely analogous. My point is that a lakebed
> develops alongside a lake, so of course they will match.
>
So what was the lake analogous to, and what was the lakebed analogous to? Or had you just provided an example that was not analogous at all?
> > >
> > > > > > > > > But why that particular feature?
> > > > > > > > >
> > > > > > > >
> > > > > > > > Because we are looking for features which can help favour certain
> > > > > > > > hypotheses
> > > > > > > > using objective bayesian inference.
> > > > > > >
> > > > > > > That's not an answer. What's special about that feature as opposed
> > > > > > > to
> > > > > > > galaxies, or matter-antimatter asymmetry, or an expanding vs.
> > > > > > > contracting vs. steady-state universe?
> > > > > > >
> > > > > >
> > > > > > Because, using objective bayesian inference, what inference regarding
> > > > > > which
> > > > > > hypotheses to favour would those other features help us with?
> > > > >
> > > > > They would not help us any more or less than the complex chemistry
> > > > > feature would.
> > > > >
> > > > > > > > using objective bayesian inference.
> > > > > > >
> > > > > > > > What other candidate feature were you
> > > > > > > > considering, and what hypotheses would it be used to distinguish
> > > > > > > > between?
> > > > > > >
> > > > > > > I've already explained my alternative to you: God is interested in
> > > > > > > black
> > > > > > > holes. The universe was created as a laboratory. Life is just some
> > > > > > > minor
> > > > > > > infestation occurring in a few tiny corners of the universe.
> > > > > > > Neither
> > > > > > > life nor complex chemistry is of any interest to God. All he wanted
> > > > > > > was
> > > > > > > a universe with gravity and enough mass to produce black holes.
> > > > > > >
> > > > > >
> > > > > > So what inference regarding which hypotheses to favour would black
> > > > > > holes
> > > > > > help
> > > > > > us with? It would not help favour that hypothesis against the
> > > > > > hypothesis
> > > > > > that
> > > > > > God exists and that we are given the experience we are given to base
> > > > > > moral
> > > > > > choices on, because the constant values which favour complex
> > > > > > chemistry
> > > > > > would
> > > > > > presumably all contain black holes, and there would be a vast range
> > > > > > of
> > > > > > parameter space in which at least one black hole would have occurred
> > > > > > and
> > > > > > yet
> > > > > > there would be no life forms.
> > > > >
> > > > > The above statement is an assumption on your part. You keep assuming
> > > > > that black holes are far more common than complex chemistry, but you
> > > > > have failed to demonstrate this.
> > > > >
> > > >
> > > > Well if you look at the
https://en.wikipedia.org/wiki/Fine-tuned_Universe
> > > > web
> > > > page you will see examples. Look at the Omega one for example:"If gravity
> > > > were too strong compared with dark energy and the initial metric
> > > > expansion,
> > > > the universe would have collapsed before life could have evolved."
> > > > Collapse
> > > > into....a black hole (I assume). Also even in universes where the
> > > > chemistry
> > > > is not complex why should black holes not be formed? Take the hydrogen
> > > > only
> > > > universe I mentioned earlier. It is formed by increasing the mass of the
> > > > down
> > > > quark by at least a factor of 3. This would cause the neutrons to decay,
> > > > even
> > > > inside nuclei, meaning only one type of atom and one chemical reaction.
> > > > But
> > > > the increase in mass would I have thought made black holes more likely,
> > > > as
> > > > gravity would I have thought led the atoms to gather together as they
> > > > would
> > > > have done in the first stars. But in this case there would be no
> > > > reactions to
> > > > raise the temperature to delay the collapse into a black hole. I would
> > > > expect
> > > > the same to be the case for those other simple universes I mentioned
> > > > (which
> > > > were all theoretically created by messing with the up and down quark
> > > > masses).
> > >
> > >
> > > I am familiar with the wikipedia examples. But I do not find them
> > > particularly compelling.
> > >
> >
> > Well even the example I gave with the up quark, down quark and electron
> > masses, where I constrained the mass consideration to a mass that a quark
> > (the mass of a top quark) has. The chances (assuming all mass values being
> > considered to be equally as likely) of randomly picking a value for each
> > which allowed complex chemistry would be about 1 in 5 trillion chance. So I
> > think it is reasonable to assume that the three of them all happening to be
> > in that range to seem like a coincidence (given that you have no explanation
> > for it). What were you thinking it required for something to seem like a
> > coincidence?
> >
> > And perhaps (up to you, but if you don't then don't just snip this bit)
> > relate that to the part (that you snipped) where I had mentioned that in the
> > book "The Grand Design" by Stephen Hawking and Leonard Mlodinow at the bottom
> > of page 206 it is written: "What can we make of these coincidences? Luck in
> > the precise form and nature of fundamental physical law is a different kind
> > of luck from the luck we find in environmental factors. It cannot be so
> > easily explained, and has far deeper physical and philosophical implications.
> > Our universe and its laws appear to have a design that both is tailor-made to
> > support us and, if we are to exist, leaves very little room for alteration.
> > That is not easily explained, and raises the natural question of why it is
> > that way."
>
> Without actually knowing what conclusions they ultimately draw, I cannot
> comment. I've not read the book as I've never particularly enjoyed
> Hawking's popular writings.
>
The conclusion that Stephen Hawking and Leonard Mlodinow drew was that reality was a multiverse.
You never mentioned what were you thought it required for something to seem like a coincidence.
> Andre
>
> --
> To email remove 'invalid' & replace 'gm' with well known Google mail service.