"Most of the fundamental ideas of science are essentially simple, and may, as a rule, be expressed in a language comprehensible to everyone." (Einstein and Infeld, The Evolution of Physics)
I had a similar question as OP about energy while going through Dave Farina's course on Classical physics (https://youtube.com/playlist?list=PLybg94GvOJ9HjfcQeJcNzLUFxa4m3i7FW).
What actually are energy and work? What features of reality are we talking about when we use these words? I don't think it's enough to say that they are useful but arbitrary definitions. And I think we can do better than saying than energy is not to be intuitively understood in and of itself. A definition is useful because it picks out some relevant aspect of nature, something real that our words correspond to. That's how our words have meaning. If we refuse to refer to intuition, we lose this meaning and just use formulas by rote. Aspects of nature can be intuited in our experience, and we use scientific concepts to represent and analyze them. So what aspects of nature are represented by work and energy? After some reflection, here are my answers to OP:
- Energy is acceleration that has been materialized and spatialized, or simply spatialized force.
- Momentum is velocity that has only been materialized. Unlike energy, the spatialization step is not carried out, nor is there a change in the velocity.
- To spatialize force, we must multiply $F$ by a spatial length, $d$. Multiplying by $t$ extends force in time instead of space. But this only cancels out one of the two divisions by $t$ that we've already done to get from displacement to velocity to acceleration. It brings us back from $ma$ to $mv$ and so $F \Delta t = \Delta mv$.
See below for the full explanation. This is basically an extended version of dimensional analysis with some interpretation. I'll first stimulate intuition using an analogy from the way kinematics becomes dynamics. Then I'll give a definition of energy from first principles using as little math as possible. Finally I'll answer the 3 questions more fully. I think the key is to use analytical imagination to form intuitive concepts of physical quantities and their combinations.
From kinematics to dynamics
The key move in dynamics is the introduction of mass as a quantity. Kinematics discusses displacement, velocity and acceleration, but it abstracts them from the matter of the objects involved. We include the dimension of mass by multiplying each quantity from kinematics by $m$:
$$d \rightarrow m \cdot d\\v \rightarrow m \cdot v\\a \rightarrow m \cdot a\\$$
This is what Newton did when he referred to momentum as a meaningful quantity measured in kilograms-meters per second that combines both the rate of motion and the quantity of matter of an object. In the same way, force takes account of both the mass and the acceleration of an object, not just the acceleration as in kinematics. I don't know if $m \cdot d$ has a name, but we could call it something like "material extension" or a "length of matter".
In effect the kinematic concepts are made more concrete by including the factor of mass, which is a concrete reality in nature that we know by intuition (i.e. by the seeing/feeling that matter has resistance). We can therefore call this procedure of introducing mass the materialization of displacement, velocity and acceleration. $d, v, a$ are abstract concepts in kinematics, and we make them less abstract by including mass alongside them. That's how we get from kinematics to dynamics.
'Spatializing' kinematics
Now let's take the above procedure, but instead of introducing the dimension of mass let's introduce the spatial dimension. We spatialize displacement, velocity and acceleration by including the spatial concept of displacement, distance or length. We do this by multiplying each by $d$. We get:
$$d \rightarrow d \cdot d\\
v \rightarrow d \cdot v\\
a \rightarrow d \cdot a\\$$
The first is a "length of a length", or simply area, measured in $m^2$. We can call the second a "length of motion" by analogy with Newton's "quantity of motion" for $mv$. Here we want to imagine a single spatial dimension that is not empty (as $d$), or filled with matter (as $m \cdot d$) but "filled" with motion ($d \cdot v$). Finally we have a "length of acceleration", $d \cdot a$, which is a space that "contains" acceleration and nothing more. Our imagination can make these abstract combinations, even though we never encounter a "length of motion" or a "length of acceleration" as separate realities in experience. Acceleration is always of some mass, in some specific context, etc. But in science we abstract away to focus on separate elements.
Work and energy
Based on the formula $W = F \cdot d$ and the above discussion we can give the following definition of work:
Work is spatialized force.
To multiply $F$ by $d$ just means to extend force in space, or to 'spatialize' it. More fully we can say that work is spatialized and materialized acceleration, which is evident after simple replacement: $W = m \cdot a \cdot d$. When work is done, acceleration is 'combined' with mass on the one hand, and with distance on the other. Work thus produces a length of force, or a length of a material quantity of acceleration. We could also say that work actualizes a force in space by taking account of the length of the space over which the force is applied.
We get to kinetic energy by working with the formulas. Assuming initial velocity of $0$ and constant $a$:
$$d = {1 \over 2}vt \text{ , } a = {v \over t}$$
Therefore spatialized acceleration from above reduces to:
$$d \cdot a = {1 \over 2}vt \cdot {v \over t} = {1 \over 2}v^2$$
Then we introduce mass to get acceleration that is both spatialized and materialized, or work, which is equal to the change in kinetic energy:
$$d \cdot a \cdot m = W = {1 \over 2}mv^2 = E_k$$
To answer OP's questions:
- Energy can be defined without circularity as spatialized and materialized acceleration, or simply as spatialized force, measured in $Nm$ or Joules. This only refers to our intuitive concepts of space, matter/mass and acceleration. (Acceleration in turn refers to the concepts of change, space and time.) It's true that we start with the formula $W=mad$, and you could say that we group m, a and d by arbitrary choice. But this grouping refers to an aspect of concrete reality, and that's what a definition expresses. It's not enough to just give symbols and logical operations. Our physical concepts actually refer to nature that is outside of us.
- Energy is acceleration that has been materialized and spatialized. Whereas momentum is velocity that has only been materialized ($mv$). The spatialization step has not been carried out, nor is velocity changing. If we spatialize momentum, we will get a length of momentum, or $mvd$. If we then take time rate of change of its velocity, we will get work or $mad$. We can say that momentum is the constant motion of a mass, whereas energy is the acceleration of a mass that has been extended in space.
- $F \cdot d$ lets us 'spatialize' force: it represents the extension of force in space. The reality of that Force-space is what we mean by energy. On the other hand $F \cdot t$ 'temporalizes' force or extends it in time. However, we've already divided by time twice to arrive from distance to velocity and from velocity to acceleration (hence the units of force are $kg \cdot {m \over s^2}$). So the t in $Ft$ will cancel one of these divisors to give us $mv$, and thus $Ft$ is impulse or the change in momentum.